A low CPM is not a bargain, it is a red flag

The same post, published organically, had produced 7.7 % engagement and a 5.3 % click-through rate. The paid version was seen by fifty times more people and left no human trace at all.

A very low cost per thousand should trigger worry, not satisfaction. In B2B, the audiences that matter are expensive because they are scarce. If the inventory is cheap, it is not that you negotiate well, it is that nobody else wants that inventory.

Where the money actually went

On a three-figure test budget split across four amplifications, the raw result was twenty-one followers gained in the first month, and none in the third. The platform's own demographics report explains why.

87 % of the budget was spent outside the target geography. No African country appeared in the top twenty-five lines of the breakdown, while the target was African. The three best-served countries were Mexico at 9.2 %, Poland at 4.4 % and the United Kingdom at 2.0 %.

Cost per follower comes out at 4.76 dollars, for zero commercial contacts. And one detail worth the trip: 145 billed clicks for twenty page visits. Most of the paid clicks were text expansions, the familiar « see more », which the platform bills as a click.

Three defaults, all wrong, all silent

The default suggested budget was twenty-eight times the intended one. The platform suggests the envelope that suits it, over a duration that suits it, and accepting it in one click is the interface's default behaviour.

The suggested audience targeted the head office country for a release about an event on another continent. The suggestion is built from the page, not from the content.

And two options were live without having been asked for: audience expansion, and the off-platform partner network. Those are what produce the volume, the derisory cost per thousand, and the nil engagement.

The price to accept, announced in advance

Tightening the geography from twenty-five countries to twelve moves the cost per thousand from 7.57 dollars to a range of 15 to 25 dollars. In practice, for the same envelope, roughly 4,000 to 6,500 impressions instead of 13,216.

That collapse in volume is the expected result, not the failure. It has to be announced before the campaign, written into the recommendation, and repeated in the report. Otherwise the client opens the dashboard, sees impressions divided by three, and logically concludes that the tightening broke the campaign.

That is the hardest part of the job, and it is not technical. It consists of getting an uglier number accepted in exchange for a better result, by someone who has been watching the ugly number for years.

What I look at instead

The real geographic split of the spend, not the split of the declared targeting. Those are two different screens and they rarely agree.

Paid engagement rate against the same content organically. If paid is far below, the money is buying the wrong eyes, and no creative adjustment will fix it.

The composition of billed clicks. A text expansion is not a visit. The gap between billed clicks and sessions that actually reached the site is the most honest measure of waste there is.

And cost per qualified follower, never cost per follower. Twenty-one followers of whom none sits in the target geography cost the price of a follower and are worth none.

A format constraint that gets expensive if found late

LinkedIn only allows four formats to be boosted: single image, video, event and document. On this file, six visuals had been produced and the post published before that limit was checked. The budget could not be placed on it.

The lesson is mundane and I repeat it anyway: eligible format is verified before the content is produced, not after it is published.