Why Ad Clicks Don't Match Website Visits

Tony CarnevaleNovember 21, 202612 min read

Venbit Intelligence Research

Tracked websites · Autumn 2026

Fewer than 2 in 100 ad clicks that reached a page were bots.

Google Ads1.8%bots

58,041 people1,060 bots

Meta1.4%bots

30,002 people416 bots

Visits that arrived with the platform's click id. Bots named by the Bot filter.

Visits that reached tracked websites carrying a Google or Meta click id: 1.8% and 1.4% were bots. Venbit Intelligence Research, 2026.

The short answer

Across tracked websites, only 1.8% of the Google ad clicks that reached a page were bots, and 1.4% of Meta's. So the gap between clicks and visits is mostly people: taps on the call button, double taps, visitors who leave before a slow page loads, blocked scripts and click ids lost in a redirect.

What to know

  • Bots made 1,060 of the 59,101 visits that arrived with a Google ad click id across tracked websites, 1.8%. For Meta's click id it was 416 of 30,418, 1.4%.
  • Phones carried 86% of Google ad-click visits and 90% of Meta's on local service websites, against 73% of all their visits.
  • A quarter of Google ad-click visits, 26%, arrived with no referrer, and an analytics tool that ignores the click id files every one of them as direct.
  • Google bills a tap on an ad's call button as a click, and that tap never opens the website, so it can never become a visit.
  • Ad platforms send their own bots as well: Google's Ads bot ran the page script 24,041 times in 90 days.
  • We see visits and never the platform's click count, so the size of a normal gap is not research. The tool's bands are rules of thumb.

Your ad platform counts clicks and your analytics counts visits, so the two numbers will not match, and bots explain very little of the difference. Venbit Intelligence Research looked at every visit that reached tracked websites over 90 days carrying a Google or Meta click id. The bot filter named 1.8% of the Google ones as bots and 1.4% of the Meta ones. Everything else in the gap is people and plumbing: taps on the call button, double taps, slow pages on phones, blocked scripts and click ids lost on the way. Enter your own two numbers below to see where yours sits.

Check your ad clicks

Check the gap between your ad clicks and visits

Pick the platform, then enter the clicks it reports and the ad visits your analytics shows for the same days. The tool places the gap in a band and lists what to check first.

Ad platform

Website clicks only: Segment, then Click type

Visits from Google Ads, the same days

Enter both numbers for the same days. The tool shows the share of clicks that never became a visit, whether that share is normal for Google Ads, and what to check first.

Open the page your ads point to in your analytics and read its sources. Works when only ads send people to that page.

% Direct

No-referrer shares: Venbit Intelligence Research, local service websites, bots excluded. The bands are our rules of thumb, not research: we see visits, never the platform's click count. Nothing you type leaves your browser.

One limit belongs at the top. Venbit Intelligence sees visits, the clicks that reached a page and ran its script, and never the platform's own click count, so the size of the gap is not something we measured. The bands in the tool are my rules of thumb, drawn wide on purpose. What we did measure is who the arriving clicks were and how they arrived, and that settles the bot question.

Under 2% of ad clicks that land are bots

I expected the bot share to be several times this size. On these websites, people outnumbered bots 55 to 1 among the visits that arrived with a Google click id, and 72 to 1 for Meta. Whatever is opening the gap on your account, the bots we can see are a sliver of it.

PlatformClick idVisitsPeopleBotsBot share
Google Adsgclid, gbraid, wbraid59,10158,0411,0601.8%
Metafbclid30,41830,0024161.4%
Microsoft Adsmsclkidn/an/an/aToo few websites
Visits that arrived with an ad platform's click id, people and bots. Venbit Intelligence Research, tracked websites, autumn 2026.

Two limits sit under those rows. The Meta row counts every visit whose address carried fbclid, and Meta's documentation describes that id for ad clicks without saying whether ordinary links carry it too. And a bot that clicks an ad and never runs a script leaves no visit for us or any other analytics to see. That kind sits inside your gap, where the platform's own filter is the one placed to catch it.

The people behind those clicks behave like customers. On the websites that got 100 or more Google ad-click visits in the window, the median website turned 3.1 of every 100 into a lead, meaning a form sent or a phone number tapped. Across the websites with that many Meta click-id visits, the median was 1.2.

55 to 1
people to bots among Google ad-click visits
3.1
leads per 100 Google ad-click visits on the median website with 100 or more
1.2
leads per 100 Meta click-id visits on the median website with 100 or more
Venbit Intelligence Research, tracked websites, bots excluded

Ad clicks arrive on phones and without referrers

86%
of Google ad-click visits came from a phone
90%
of Meta click-id visits came from a phone
26%
of Google ad-click visits arrived with no referrer
6%
of Meta click-id visits arrived with no referrer
Venbit Intelligence Research, local service websites, bots excluded

Across all visits to the same websites, 73% came from a phone, and ad clicks lean further still, which matters for the gap. The platform counts a click the moment someone taps. Your analytics counts a visit only after the page has loaded far enough for its script to run, and on a phone over cell data that wait is where people give up. The guide to phone and computer visitors covers how the two behave once they arrive.

The no-referrer figure surprised me more. A quarter of the Google ad clicks arrived with nothing to say where they came from except the click id on the address, and the data does not say why each one lacked a referrer. An analytics tool that reads the referrer and ignores the click id files every one of those visits as direct, so the ad report comes up short while direct looks busy. Meta's figure is far lower at 6%.

Landing page

/care-plans?gclid=Cj0KCQjw…

Referrer
None
Click id
gclid · Google Ads
Clicked an adPlatform google

Traffic sources

Ads1,787
Direct1,362

Research: 26 in 100 Google ad-click visits arrive with no referrer, across tracked local service websites.

A Google ad click lands on the demo site with no referrer. The tracker reads its gclid, the trail says Clicked an ad, and Traffic sources counts it under Ads instead of Direct.

If your ads run and your direct traffic climbs in step with them, this is the first place I would look. The guide on direct traffic covers everything else that lands in that bucket, from typed addresses to links in text messages.

Six reasons ad clicks and visits disagree

These are the causes in the order I look for them on a small business account. Google's own help page on clicks and sessions lists most of them, and I keep to its wording where it has one.

  • The call button counts as a click

    A tap on the call button in a Google ad is billed the same as a headline click, and it dials you without opening the website. The Clicks column adds the two together; segment it by Click type to split them.

  • They leave before the page loads

    Someone clicks, then closes the browser or moves on before the analytics code loads, which Google's help names as a cause. A slow page on cell data turns paid taps into clicks with no visit.

  • One person taps twice

    A second tap on the same ad is a second click. Google's help says Analytics can record two clicks from one person within 30 minutes as one session.

  • A blocker stops the script

    Ad blockers and privacy browsers stop many analytics scripts from loading, so the visit happens and nothing records it. Plausible found 58% of visitors blocking Google Analytics on one site in 2021, a technical audience far from a homeowner on a phone.

  • The click id is lost in a redirect

    Redirects on a landing page can keep the analytics code from running and strip the campaign details from the address, Google's help warns. The visit then counts under direct or organic, away from the ad.

  • Invalid clicks are already refunded

    Google filters the clicks it judges invalid, leaves them out of the Clicks column and does not charge for them. Invalid activity found after billing comes back as a credit. Its help says there is no industry standard for a normal invalid traffic rate.

Only the first five push clicks above visits. The sixth works the other way, because a filtered click that loads the page still reaches your analytics as a visit. For a local service business the first two matter most: ad visits here were 86% phones, and on a campaign built to make the phone ring, every tap on the call button sits in the Clicks column with no visit behind it.

Ad performance
Last 30 days · Sep 21 to Oct 21 · Needs attention
The retargeting ad lands people on the homepage.
See more
79%
of retargeting clicks leave within five seconds
All ad clicks by day, amber became leads
Oct 14 · 4 leads
Sep 22Oct 21
Next step

Send it to /pricing, where the price it advertised is the first thing on the screen.

The Feed's ad post on the demo site: 79% of the retargeting clicks left within five seconds of landing, the visible edge of the people who leave before a page has loaded.

Venbit Intelligence counts a visit once the page's script runs, so the people who leave a second later appear as Left within 5 seconds in the Ads report. Those who leave before the script runs appear nowhere, in any analytics. A high Left within 5 seconds and a wide gap on the same campaign point at the same landing page.

When visits outnumber ad clicks

The gap can run the other way, and it has causes of its own. Google's help describes a visitor who bookmarks the landing page with the gclid still on the address: each return reads as an ad visit with no new click. Clicks Google filtered as invalid also reach your analytics as visits.

CheckerSent byVisits
Google Ads botGoogle24,041
OAI-AdsBotOpenAI370
Meta ad reviewMeta, inferred from how the visit behaved271
Ad platforms' own checkers that opened landing pages and ran the page script. Venbit Intelligence Research, tracked websites, over 90 days. Kept out of every figure on this page.

Then come the platforms' own bots. Google's Ads bot opens the landing pages behind a Google Ads campaign, and it ran our script 24,041 times in 90 days, on about four in five tracked websites. That is about 41 visits for every 100 people who arrived from a Google ad. Our bot filter names it and keeps it out of the count, and an analytics tool that does not filter it counts those visits as people.

The guide to bot traffic covers the other kinds and has a tool that sets your host's count beside your analytics. If the clicks look fine and the leads look fake, how to tell if a website lead is spam covers the form side.

Find where your missing clicks went

  1. 1

    Compare the same days in the same time zone

    Set both reports to the same dates. If the ad account and the analytics keep different time zones, an evening's clicks land on the next day's visits and make a gap out of nothing.

  2. 2

    Take the call button taps out of the click count

    In Google Ads, open Segment, then Click type, and keep only the clicks that went to the website. Meta reports Link clicks apart from Clicks (all), and Link clicks is the figure to compare.

  3. 3

    Open the landing page on a phone with Wi-Fi off

    Count the seconds before anything appears on the screen. Every second on a blank page is time for a paid visitor to give up before the analytics runs.

  4. 4

    Check that the click id survives

    Paste the ad's final address into a browser with ?gclid=test on the end, or fbclid=test for Meta, and read the address bar once the page settles. If the tag is gone, a redirect dropped it, and those visits are being filed under direct or organic.

  5. 5

    Find your own visit in the real-time report

    Open the landing page and watch your analytics' live view. If your visit never appears, the script is missing from that page or is waiting on a cookie banner you have not answered.

  6. 6

    Read the Invalid clicks column

    Add it to your Google Ads campaign table. Those clicks were filtered and not charged, so they are already out of the Clicks figure you are comparing.

If the clicks and visits agree and the conversions do not, that is a separate question with its own answer: why Google Ads conversions don't match analytics. That article covers conversion dates and view-through conversions, which this one leaves alone.

See every ad click that reached your site

Venbit Intelligence counts the ad clicks that reached your site from the moment each one lands. The tracker reads the click id on the landing address and keeps it with the visit, so a Google or Microsoft click is filed as paid with no tags on the link and no ad account to connect. A Meta click carrying only fbclid is filed under Meta as social, and moves to Ads once the ad's link is tagged as paid.

venbit.com/dashboard
Ad performance
Yesterday1D1W1M3M6M1YAll
1,777ad clicks▲ 14%
48 leads9 more on a later visitLast 30 days
Leads
48+26%
38 previous period
Lead rate
2.7%+12%
2.4% previous period
Time on site
1:12
1.8 pages a visit
Stayed to read
47%
38.4% rest of site
Left within 5 seconds
41%
985 clicks
Started the form
8.6%
13.9% saw a form
Clicks by day
Sep 29Oct 14
The clicks that became leads
Platforms
PlatformClicksLeadsLead rateStayedLeft in 5s
Google Ads2 campaigns1,173443.8%45.2%43.1%
Meta1 campaign60440.6%14%79%
Campaigns
CampaignChangeClicksLeadsLead rateStayedLeft in 5s
Care plans - exactGoogle Ads▲ 14%924384.1%47%41%
Retargeting - site visitorsMeta▲ 14%60440.7%14%79%
The Ad performance tab on the demo site: 1,777 ad clicks that reached the site and 48 leads in 30 days, with clicks, leads and the share that left within 5 seconds for each platform.

The Ads report gives each platform and each campaign its clicks, leads and lead rate, set against the period before. It never sees the platform's own click count or its billing, so its clicks sit on the same side of the gap as your analytics, and a wide gap is something you read off your ad account beside it.

Every lead in the Inbox carries the ad that brought it. A form sent from a Meta click reads Meta ad and the campaign name, with the page, the pages read, the town and the device, so the count of leads per campaign comes from your own site's records. The lead source guide shows how every lead keeps its source, paid or unpaid.

Inbox2ArchivedSpam 2Deleted
/Sent the form · Minster, OH · Website Maintenance Campaign 1
/checkupSent the form · Dallas, TX · Website Maintenance Campaign 1
/Sent the form · Crockett, TX · Website Maintenance Campaign 1
/Sent the form · Miami, FL · direct
/Sent the form · Belleville, IL · direct
A lead from Minster, OH opened in the Inbox on the demo site, with the Meta ad and campaign it came from, the page, the pages read, the town and the device.

The bot filter runs before anything is counted. Named crawlers, the ad platforms' own checkers and visits that behave like scripts are set aside, so the ad clicks in the report are the 98 in 100 that were people.

Bot filter · Kept out of your numbers

Last 30 days · Sep 21 to Oct 21 · Worth knowing

6,447 visits from bots were kept out of your numbers.

That is 43% of everything that reached your site. Uptime monitors made 4,233, and AI crawlers made 949. Every figure on this page leaves these out.

8,510people · 57%

6,447bots · 43%

Uptime monitors4,233

AI crawlers949

Search engines and ad checkers853

SEO tools338

Link previews74

How we know
The Feed's Bot filter post on the demo site lists the 6,447 bot visits kept out of the count for 30 days, 853 of them search engines and ad checkers.

How we measured ad clicks and bots

The figures come from Venbit Intelligence Research: websites tracked in the Venbit Approved System, 21 August to 18 November 2026. A visit counted as an ad click when its landing address carried the platform's click id: gclid, gbraid, wbraid, dclid or gad_source for Google, fbclid for Meta, msclkid for Microsoft. Bots were named by the same filter that keeps them out of every report, from their user agents and from how the visit behaved.

The phone and referrer shares come from the local service websites in the set, with software, content and B2B sites left out. Microsoft Ads ran on too few of the websites to publish a figure. The bot share counts bots that ran the page script; crawlers that read a page without running it are counted elsewhere, and script-based analytics never shows them as visits either.

These are websites whose owners chose to track them closely, and about half of them buy enough ads to count as advertisers. Read the figures as a well-measured sample of businesses like these, and read the Meta row knowing its click id also rides on posts nobody paid for.

See every ad click that reached your site

Venbit Intelligence reads the click id on every ad visit, keeps the bots out and names the campaign on every lead. One line to install, free for 14 days, no card needed.

In Venbit Intelligence

More website analytics guides

Every question in this series, from Website Analytics for Small Business Owners.

Leads inbox4
Ad performance4
SEO tracking3
Visitor behaviour4
Feed and notifications3
Goals and measurement4
Privacy, AI and teams3
Venbit Intelligence Research19

Questions & Answers

  • Google counts the tap, and your analytics counts the page that loaded after it. Taps on the call button, double taps, people who leave before a slow page loads on a phone, blocked scripts and click ids lost in redirects all sit in between. Bots are a small part: 1.8% of the Google ad-click visits we measured across tracked websites.

  • On the small business websites we measured, 58,041 of the 59,101 visits that arrived with a Google click id were people, 98.2%. The median website with 100 or more of those visits turned 3.1 of every 100 into a lead. Google also filters invalid clicks before billing and leaves them out of the Clicks column.

  • Very few of the ones that reach your site. Of 30,418 visits carrying Meta's click id across tracked websites, our bot filter named 416 as bots, 1.4%. Meta adds that id to ordinary Facebook and Instagram links as well as ads, so the figure covers both. A bot that never loads the page is invisible to every analytics tool.

  • Check four things in order: the analytics script runs on the landing page, no redirect strips the gclid, auto-tagging is on, and nothing holds the script back until a cookie banner is answered. On the local service websites we measured, 26% of Google ad-click visits arrived with no referrer, so a tool that drops the click id files them as direct.

  • Nobody can measure that from the website side, because a site never sees the platform's click count. Our rule of thumb is up to 20% for Google and Microsoft and up to 30% for Meta, whose clicks come from a feed on a phone. Google's older Analytics troubleshooter asks whether the gap is under 20% in most campaigns.

  • The visit arrived with no referrer and your analytics did not read the click id. That was the case for 26% of Google ad-click visits on the local service websites we measured, against 6% for Meta. Turn on auto-tagging, keep redirects from stripping the gclid, and use analytics that reads it.

  • Google does not charge for invalid clicks it catches before billing, and invalid activity found after billing comes back as a credit. Add the Invalid clicks column to see what was filtered. Its help also says there is no industry standard for a normal invalid traffic rate, so judge yours against your own months.

More guides with a tool

See all 23 guides