Billed at the Tap, Counted at the Arrival
A billed click and a recorded visit are two different events, separated by a network, a browser and a few seconds of human patience.
Meta counts a Link Click the moment someone taps the link in your ad. Whether the destination page loads is not part of the definition. Meta also reports Landing Page Views, which counts only the clicks where the destination actually finished loading and your pixel fired. Both metrics live in the same report, on the same rows, and you have been scrolling past the second one for years.
Google says the same thing about its own numbers, in plainer language. Its help documentation for the clicks and sessions discrepancy states that clicks and sessions are different metrics and that some level of discrepancy is expected. That sentence is doing a lot of work. It is the platform telling you, in advance, that the number it bills on and the number you measure on were never designed to be the same number.
Every reading in the middle column is reasonable, and every one of them skips the several seconds where the money actually goes missing. The 1,000 and the 780 are illustrative arithmetic, not a measurement of any account.
The Cheapest Measurement You Are Not Taking
Meta's own documentation draws the line for you. A Link Click is a click on a link in your ad. A Landing Page View is counted when that click results in the destination page successfully loading and the Meta Pixel firing. All Landing Page Views are Link Clicks. Very few campaigns see the reverse hold true.
That is the cleanest available measurement of the arrival gap on any platform, and it is already sitting in your account. You do not need a vendor, a tag audit or a consultant to see it. You need one report, two columns and about ninety seconds.
The reason so few people look is that neither number is presented as suspect. Ads Manager shows both with equal confidence, in the same font, on the same row. There is no warning label on the smaller one saying "this is the part that survived the trip".
Where the Click Dies Between the Ad and Your Log
The gap is not one leak. It is a sequence of them, and they happen in a fixed order along the path from the tap to your first log line.
Only the two teal steps are yours to change. The rest you can measure, plan around, and stop mistaking for audience quality.
Start at the left. The billing rule is fixed and you will not argue Meta out of it.
The second step is yours. Every hop between the ad and the offer is another chance to lose the visitor: a DNS lookup, a TLS handshake, a slow tracker, a network's own redirect, a cloaker. Google's documentation warns that redirects on a landing page can lose the campaign information appended to the URL, including the click ID, which means a hop can cost you the visitor and the attribution separately. On a phone, on a poor connection, people simply leave.
The third is partly yours. A page-load-triggered tracking pixel needs the page to reach the point where the script executes. Google states it directly: a user who navigates away before the tracking code fully loads produces a click and no session. Someone who taps back at second two has been billed and will never be counted.
The fourth is not yours at all. Content blockers and browser protections stop vendor scripts running, and a refused consent banner does the same job by a different route. The server received the request. The tag never ran. Google's help page lists this under browser preferences that prevent Analytics from reporting on those users while the ad platform still reports them.
The fifth is a counting rule rather than a loss. Google spells it out: two clicks on the same ad inside 30 minutes without closing the browser may register as a single session, while Google Ads counts two clicks. Sessions structurally undercount clicks even with perfect delivery.
The 1,000 and the 780 Are Arithmetic, Not a Finding
The numbers in the headline are an illustration of the shape, not a measurement of anyone's account. I picked them because they make the ratio legible at a glance: bill on 1,000, count 780, and you are computing every downstream figure on 78% of what you paid for. I did not run a study across advertisers and I am not going to pretend otherwise.
I am also not going to quote you an industry average, because nobody publishes one that survives scrutiny. Search for a benchmark ratio of billed clicks to recorded arrivals and you will find plenty of confident percentages with no sample, no method and no date attached to any of them. Google's own material says only that some level of discrepancy is expected. That is the most specific honest thing anybody has published.
One number I will quote, because it has a named publisher: GWI's figures, as reported in DataReportal's Digital 2026 Global Overview Report, put ad-blocker use at 29.5% of internet users worldwide as of Q2 2025 and 32.5% in the United States. Blockthrough, on the same subject, puts US penetration at 27% on desktop and 22% elsewhere. Those two sets of figures do not agree, which is precisely why you should measure your own audience instead of adopting either.
The Gap Runs In Both Directions
This is where most people get the diagnosis backwards, so it is worth being exact.
Google removes invalid clicks from your Ads reports and does not charge you for them. It does not remove the resulting sessions from GA4. So invalid traffic pushes your session count up while pushing your billed click count down, and the two systems disagree in the opposite direction from everything above.
There are more reversals. If someone bookmarks your URL with the click ID still attached and returns later, Analytics records a session from your campaign that you were never billed for. A returning visitor inside a campaign's lifetime does the same. Google's documentation says that having more sessions than clicks is an indication of positive engagement, which is true, and also means the sign of the discrepancy tells you almost nothing on its own.
So the arrival gap is not a single subtraction. It is a net figure with losses on one side and phantom gains on the other, and the only way to read it is to hold the three numbers separately instead of assuming any one of them is the truth.
The Damage Runs Sideways, Not Downwards
If the gap were uniform, it would be a nuisance and nothing more. Everything would be off by the same factor, your rankings between campaigns would hold, and you would carry on.
It is not uniform. Slow mobile connections lose more clicks before the page loads. Blocker-heavy audiences lose more tags. Some placements route through more hops than others. The loss correlates with exactly the things that define a segment.
Which produces the failure that actually costs money. The segment losing the most on arrival looks like your worst-converting segment, because the conversions it produced were divided by a visit count that never recorded them. You cut budget on traffic that was arriving fine and converting fine and simply not reporting. Then you shift spend toward the segment with the cleanest measurement, which is not the same thing as the segment with the best economics.
That failure has no ceiling. It compounds every time you act on it, because each round of cuts pushes more of your budget toward whatever is easiest to measure.
This is a different animal from the one in the GA4 measurement piece. That was about the conversion event going missing at the point of sale. Here the conversion may be recorded perfectly. What went missing is the visit it should have been divided by.
Three Numbers, One Campaign
Three questions, and the first one is free. Most accounts stop at the first and never learn what the other two would have shown.
What Raises the Floor
Nothing in the ad account closes this. Meta will keep billing the tap, blockers will keep blocking, and people on bad connections will keep leaving.
What raises the floor is a server-side click log written the moment the request hits your redirect endpoint, before any browser script has had a chance to run or not run. That record exists whether or not the page loads, whether or not a blocker is active, whether or not the visitor accepts a consent banner for analytics. It is the same discipline as server-side tagging, applied one step earlier in the chain, at the click rather than the conversion.
Then reconciliation, which is the part people skip. Hold billed clicks, server-logged arrivals and tag-fired pageviews as three separate numbers. Do not average them. Do not pick a favourite. The differences between them are the diagnostic: billed minus server-logged tells you what died on the network, server-logged minus tag-fired tells you what died in the browser, and the split between those two is what decides whether you fix your redirect chain or your page.
Keep the redirect chain short while you are there. Every hop you remove is a loss you stop paying for, and it is the only part of this you can fix by yourself this afternoon.
So What Do You Do About It
Open Ads Manager tonight. One campaign. Put Link Clicks and Landing Page Views in the same view and compute the ratio, then break it down by placement and by device. Then pull your tracker's click count for the same campaign and the same dates and lay it alongside. Three numbers, one campaign, one sitting. The spread between them is your arrival gap, and until you have looked at it, every CPA you have quoted to a partner or a client was measured against a denominator you never verified.
Then decide whether the number you record deserves to depend on a browser script, given that GWI puts ad-blocker use at 32.5% of US internet users. That is the design decision behind ClickerVolt: the click is written server-side at the redirect, at the edge, before the browser gets a vote, so the arrival exists in your own records even when the page never finishes loading. See how the server-side click log works.
Whatever you run, hold on to this. The platform bills you for an intention and your tracker records an outcome, and you have been treating those as the same number. They never were.
The 1,000 and 780 figures in the headline are an illustration of the arithmetic, chosen to make the ratio readable. They are not a measurement of any account, campaign or advertiser. Vendor behaviour described here was verified against Meta and Google documentation in August 2026, and the ad-blocker figures are attributed to their named publishers rather than presented as our own research.
