What is Link click?

A campaign spends A$3,600 and reports 12,000 clicks and 8,000 link clicks. The all-clicks price is 3,600 ÷ 12,000 = A$0.30; the link-click price is 3,600 ÷ 8,000 = A$0.45. A third of the counted clicks produced no traffic, and an account comparing its A$0.30 against another account's A$0.45 is comparing two reporting settings rather than two campaigns.

Platforms that report an all-clicks figure also report a link-clicks figure, and the difference between them is the whole reason this term exists.

The wide count includes every interaction the platform treats as a click. Reactions and likes. Comments and replies. Shares and reposts. Saves. Taps on the advertiser's page or profile name. Expansions of an image, a carousel card or a truncated caption. On some placements, interactions with elements of the ad unit that lead nowhere at all.

The narrow count includes only the interactions that sent somebody to a destination.

Both are divided into spend to produce a figure called cost per click. Both are divided into impressions to produce a figure called click-through rate. Neither metric name carries any indication of which count produced it.

The gap is routinely substantial: a large minority of the wide count consists of interactions that produced no traffic. It varies by format and by creative: carousels generate expansions, video generates in-unit interactions, and an ad that provokes discussion generates comments.

The consequence is a persistent, invisible error in comparisons. An advertiser looking at a wide-count cost per click of thirty cents and reading elsewhere about fifty-cent clicks concludes they are doing well. Both figures may describe identical performance with different denominators.

The same error appears inside a single account when one campaign's reporting is configured differently from another's, and when a report built months ago is compared with one built today.

Three reasons, and they are practical rather than philosophical.

It corresponds to something. A link click is supposed to produce a page load on the destination, which means it can be reconciled against analytics. A reaction cannot be reconciled against anything.

It travels. Other platforms' click counts are closer to this definition than to the wide one, which makes cross-platform comparison approximately honest instead of systematically wrong.

It cannot be inflated by engagement. An ad that provokes a long comment thread will show a falling cost per click on the wide count while sending no additional traffic at all.

Setting the narrow figure as the account default, once, removes an entire class of misreadings permanently.

The gap between the two counts is itself informative, and tracking it costs nothing.

A campaign where the wide count is close to the narrow one is producing ads that people click through rather than interact with in place — typical of direct-response creative with a clear offer.

A campaign where the wide count is several times the narrow one is producing ads that people respond to without leaving. Sometimes that is the point, as with a brand campaign or a post being seeded for social proof. More often it is a diagnosis: the ad is entertaining, or provocative, or complete in itself, and it gives nobody a reason to go anywhere.

A widening ratio over the life of a campaign usually means fatigue has set in on the traffic side before it set in on the interaction side — people still recognise and react to the ad, and they have stopped wanting what is behind it.

Some platforms separate these further. A link click can lead to a destination inside the platform — a shop, a profile, an in-app experience — while an outbound click leaves for an external site.

Where the distinction is reported, the outbound figure is the one that corresponds to a visit on property the advertiser controls, and it is the one to reconcile with analytics. An account sending traffic to an in-platform shop will see a healthy link-click count and very few sessions, and nothing is wrong.

Where it is not reported, the link-click figure carries both and the reconciliation has to allow for it.

Even on the narrow count, clicks and sessions will not match, and the gap has ordinary causes: people abandoning slow pages, redirects losing parameters or timing out, privacy settings blocking analytics scripts, and impatient double taps producing two clicks and one session.

A shortfall of a fifth to a third is unremarkable. What matters is stability. A gap that suddenly widens points at the destination — a deployment, a new redirect, a page-speed regression, a consent banner change — rather than at the campaign, and looking for it in the ad account wastes a day.

Reconciling the wide count against sessions, by contrast, produces an alarming and meaningless gap, because most of the numerator was never going to produce a session.

It is not useless, and it is not a click count. It is an engagement count, and it belongs beside engagement rate and cost per engagement where it is read for what it is: an indication of whether the creative earned a response from the people it reached.

Labelled that way, in its own column, it stops corrupting the traffic figures and starts being informative. Left labelled as clicks, it makes every derived cost and rate in the account optimistic by an amount that varies by creative and by placement, which is the worst kind of error — not a constant one that can be adjusted for, but a moving one that changes with exactly the things being tested.

• Set the narrower figure as the account default and leave it there. Every comparison afterwards becomes honest without anyone having to remember a caveat.
• Track the ratio between the two counts as a creative signal. A widening gap means the ad is earning interaction without earning traffic.
• Distinguish link clicks from outbound clicks where the platform separates them. An in-platform destination and an external one are not the same visit.
• Reconcile link clicks against destination sessions rather than against the wider count. Only one of them is supposed to produce a page load.
• Use the narrow figure in every cross-platform comparison, because it is the definition closest to what other platforms mean by a click.