Meta Ads: 39% of paid clicks never fired a landing page view. With a working pixel it was still 35%
Across the 20 accounts in this audit that recorded both actions, Meta counted 733,847 link clicks and 445,245 landing page views. The difference, 288,602 clicks, is 39.3% of everything paid for. The pooled lifetime spend across the 23 accounts is $783,518 (AUD), and we apportion $443,868 of it to clicks that did not land. That last figure is a proportional estimate, not a measured and calculated loss; more on it below.
A gap between link clicks and landing page views has two possible causes, and they call for different responses. Either the person left before the page loaded (a landing page problem), or the page had no Meta pixel to report that it loaded (a measurement problem).
Accounts where the pixel demonstrably fires, meaning at least one offsite_conversion.fb_pixel_* event exists anywhere in the account, still lost 34.9% of their clicks before a landing page view. Accounts with no pixel events at all lost 94.8%. The first is about landing pages. The second is about tracking.
The three actions and what each one proves
Action type | What Meta counts | What it needs |
|---|---|---|
| A paid click that leaves Meta for a destination URL | Nothing on the destination |
| The pixel's PageView fired on the destination after that click | The Meta pixel on that page, and the page loading far enough to run it |
| Standard and custom pixel events (lead, purchase, view content and so on) | The pixel, plus an event configured on the site |
A link click is counted on Meta's side of the boundary. A landing page view is counted on the advertiser's side and needs the pixel to run. So a large gap with no pixel events says the site is not reporting, not that people bounced. A large gap with plenty of pixel events says something happened between the click and the page rendering.
We used any pixel event at all as the evidence test, because it proves the pixel is installed somewhere. Accounts with zero such events are labelled "no pixel events". Accounts with a handful (three to twelve) are "weak".
Group | Accounts | Link clicks | Landing page views | Drop-off | Spend behind lost clicks |
|---|---|---|---|---|---|
Pixel firing | 11 | 651,427 | 424,001 | 34.9% | $393,158 |
Weak | 4 | 35,322 | 18,804 | 46.8% | $19,697 |
No pixel events | 5 | 47,098 | 2,440 | 94.8% | $31,013 |
Pooled | 20 | 733,847 | 445,245 | 39.3% | $443,868 |
The pixel-firing group holds most of the clicks in the table, so the pooled 39.3% is mostly their 34.9%. The no-pixel group holds a small share of the clicks and almost none of the landing page views. Three accounts (I, K and O, all professional services) are not in the table because they did not record both actions.
The account-by-account view
Acct | Band | Link clicks | Landing page views | Drop-off | Pixel events | Group |
|---|---|---|---|---|---|---|
W | Health & care services | 25,346 | 84 | 99.7% | 0 | No pixel events |
Q | Health & care services | 9,799 | 512 | 94.8% | 0 | No pixel events |
M | Home improvement & trades | 34,662 | 1,903 | 94.5% | 614 | Pixel firing |
E | Home improvement & trades | 833 | 50 | 94.0% | 12 | Weak |
H | Health & care services | 9,972 | 1,097 | 89.0% | 0 | No pixel events |
D | Professional & B2B services | 3,792 | 876 | 76.9% | 5 | Weak |
V | Home improvement & trades | 8,871 | 2,104 | 76.3% | 3 | Weak |
R | Manufacturing & equipment | 1,463 | 364 | 75.1% | 0 | No pixel events |
N | Professional & B2B services | 12,136 | 3,156 | 74.0% | 56 | Pixel firing |
G | Home improvement & trades | 39,773 | 15,785 | 60.3% | 10,666 | Pixel firing |
C | Manufacturing & equipment | 64,932 | 26,234 | 59.6% | 100 | Pixel firing |
J | Home improvement & trades | 6,745 | 2,834 | 58.0% | 134 | Pixel firing |
P | Home improvement & trades | 52,251 | 23,251 | 55.5% | 1,575 | Pixel firing |
L | Consumer goods & food | 8,028 | 3,829 | 52.3% | 1,136 | Pixel firing |
B | Manufacturing & equipment | 21,826 | 15,774 | 27.7% | 4 | Weak |
U | Health & care services | 518 | 383 | 26.1% | 0 | No pixel events |
S | Consumer goods & food | 45,051 | 35,106 | 22.1% | 38,755 | Pixel firing |
T | Consumer goods & food | 23,184 | 18,590 | 19.8% | 5,621 | Pixel firing |
F | Manufacturing & equipment | 167,014 | 134,059 | 19.7% | 130,129 | Pixel firing |
A | Consumer goods & food | 197,651 | 159,254 | 19.4% | 354,599 | Pixel firing |
The top of the table is the measurement half. Account W paid for 25,346 clicks and recorded 84 landing page views, with no pixel events. Account Q, 9,799 clicks and 512 views. Account H, 9,972 and 1,097. All three are health services, and together with account R (manufacturing, 1,463 clicks) and account U (health, 518 clicks, below any sensible floor) they make up the no-pixel group. Nobody should read W's 99.7% as a bounce rate. For almost every ad it ran, nothing on the destination could say the page loaded.
The bottom of the table is the landing page half. Accounts A, F, T and S are the four largest pixel-firing accounts by clicks, with pixel events in the tens or hundreds of thousands, and they lose 19.4% to 22.1% of their clicks. That is the gap when the pixel is beyond doubt: one click in five still never produces a page view.
In between are the accounts that make the 34.9% average. Home improvement accounts G, J and P sit at 55.5% to 60.3% with the pixel clearly firing (10,666 events for G). Account C loses 59.6% on 64,932 clicks with 100 pixel events, enough to prove installation and not enough to prove it is on every page.
Why this happens
When the pixel is firing and the gap is still a third of clicks, the causes are on the page and in the placement.
The first is the phone. Post 16 found 94.5% of the $783,518 went to mobile. A tap opens the destination in Meta's in-app browser, and a page that takes a few seconds to render on a mobile connection is one many people leave before the pixel's PageView has run. Meta only counts the landing page view once that script executes.
The second is what the click was. Post 20 found 216 campaigns choosing the LINK_CLICKS objective in 2023 and 2024. That objective tells Meta to find people who click, and it does, including people who click by accident. The system is not asked to find people who wait for the page. Placements built around fast scrolling and small tap targets add to it; the Audience Network and Reels posts earlier in this series cover click quality on those surfaces.
The third is the destination itself. A landing page view only fires on a page that carries the pixel. Ads pointing at a booking system, a third-party form, or a page the developer never tagged will register a link click and nothing else. Account M is the example: 614 pixel events say the pixel exists on the site, and a 94.5% drop-off on 34,662 clicks says the ads were mostly sending people somewhere it was not. That is why "pixel firing" in the group table means "installed somewhere", not "installed everywhere the ads go".
For the no-pixel group, the explanation is simpler. There is nothing in the account to prove the pixel exists beyond the 2,440 landing page views themselves, which need a PageView to fire. Four of the five accounts are health services. Whatever the reason (a site never tagged, a booking platform the pixel cannot run on, or restrictions on what health-category sites can send), Meta had no signal from the destination, and every optimisation decision in those accounts was made on clicks alone.
What it does to decisions
The most direct effect is on the cost of a visit. CPC is spend divided by link clicks. Cost per landing page view is spend divided by the clicks that produced a page. In the pixel-firing group, 34.9% of clicks disappear between those two denominators, so the price of a person who actually arrived is roughly half again what the CPC column says. In an account losing 60%, it is two and a half times. Every comparison in this series that used CPC (post 15 among them) has this gap underneath it.
The second effect is on what the system optimises for. Meta's Traffic objective lets you optimise for link clicks or for landing page views. The second is the one that pays for people who arrive. In the pixel-firing accounts losing 55% to 60%, that switch would change who the ads are shown to, and it costs nothing.
The third effect is the $443,868. The pull works it as a proportional estimate: each account's spend is apportioned to its clicks in the same ratio as its clicks that did not land, and the results are summed. It is not 39.3% of $783,518, because it is worked account by account and the accounts losing the most clicks are not the ones with the cheapest clicks. Nor is it money wasted in any simple sense: some of those clicks saw the page before the pixel ran, some were accidental taps that would never have converted, and $31,013 of it sits in accounts where the pixel could not have reported anything. Read it as the scale of the thing, not a refund owed.
Reading | What it would mean | Where the data supports it |
|---|---|---|
The gap is a bounce | People left before the page rendered | Pixel-firing group: 34.9% on 651,427 clicks |
The gap is a measurement failure | Nothing on the destination could report a load | No-pixel group: 94.8% on 47,098 clicks |
Both, inside one account | Pixel on some pages, ads sent to others | Account M: 614 pixel events, 94.5% drop-off |
Where it varies
Band | Accounts in the table | Drop-off range | Evidence |
|---|---|---|---|
Consumer goods & food | 4 (A, T, S, L) | 19.4% to 52.3% | All four pixel firing |
Manufacturing & equipment | 4 (F, B, C, R) | 19.7% to 75.1% | Two firing, one weak, one none |
Home improvement & trades | 6 (P, J, G, V, E, M) | 55.5% to 94.5% | Four firing, two weak |
Professional & B2B services | 2 (N, D) | 74.0% to 76.9% | One firing, one weak |
Health & care services | 4 (U, H, Q, W) | 26.1% to 99.7% | None have pixel events |
Consumer goods is the clean band: every account has the pixel firing and three of four sit at 19.4% to 22.1%. Health services is the opposite: not one of the four accounts records a pixel event, and the three with meaningful click volume lose 89.0% to 99.7%. Home improvement is where the landing page argument lives: four accounts with the pixel firing, and the lowest drop-off among them is 55.5%.
Floor notes. Account U's 26.1% rests on 518 clicks, account E's 94.0% on 833, and account R's 75.1% on 1,463. We would not build anything on those individual figures; they are in the table because they belong to the pooled groups.
What to check in your account
In Ads Manager, add the columns Link clicks, Landing page views and Cost per landing page view, at campaign level, for the last 90 days. Divide landing page views by link clicks. In this set the pixel-firing benchmark is 65.1% (a 34.9% drop-off); the best accounts are above 80%.
If the rate is under 10%, stop reading it as a landing page result. Open Events Manager and check whether the pixel has recorded any event in the last 30 days. If it has not, that is the finding, and everything else in the account has been optimised on clicks alone.
If the pixel is active, use Test Events with the exact destination URL from each ad, including its tracking parameters, on a phone.
List every destination URL in the account and mark the ones that are not on your own domain. Those will never fire a landing page view. Decide whether to accept that or to route them through a tagged page.
Break the remaining campaigns down by placement and by device. If the gap concentrates in one placement, the fix is a placement exclusion, not a page rebuild.
On any Traffic campaign optimised for link clicks, change the optimisation to landing page views and compare cost per landing page view a month later.
Test the page on a mobile connection. The pixel's PageView fires on load; if the page takes long enough that people leave first, no amount of tracking work will close the gap.
The principle
A click is counted where the platform can see it and a page view is counted where you can. The gap between them is the first place the platform's numbers and yours diverge, and it is the last place most people look. Before treating it as a landing page problem, establish whether your side of the boundary is reporting at all; a 95% gap with a silent pixel is a tracking finding, and a 35% gap with a loud one is a page finding. Only one of them is fixed by changing the page. It is where we start in our Facebook Ads management, because every metric further down the funnel inherits whatever is wrong here.