Paid social judged on what it actually caused.

Paid social has an expectations problem

It usually gets diagnosed as an attribution problem, which sends people hunting for a fix in the wrong place.

When Meta reports a conversion, it is answering a specific question: did someone who saw or clicked this ad go on to convert within the attribution window. That is a real question and the answer is usually correct. It is just not the question you are asking, which is whether the campaign caused revenue you would not otherwise have received. A platform grading its own homework will always find more of its own homework worth grading.

This is not a scandal and it does not mean the numbers are fake. It means the reported figure needs translating before it goes in front of a board. I would rather set that expectation in week one than have an awkward conversation in month four about why platform-reported ROAS and the revenue in your accounting system have never once agreed.

Judge social on last-click and you will switch off the thing that is working

This is the most expensive mistake I see, and it is easy to make because the reporting encourages it.

Upper-funnel social often does its work invisibly. Someone sees a video on Tuesday, does nothing, and searches your brand name on Saturday. Last-click hands that conversion to brand search, which then looks like your most efficient channel, because it is harvesting demand that something else created. Cut social because its last-click numbers look weak, and brand search volume starts sliding a few weeks later. By the time anyone connects the two, the budget has already been moved.

The way out of this is a deliberate test. Pause a defined slice of social, hold everything else steady, and watch branded search volume and direct traffic over the following weeks. It costs a little in the short term and settles the argument permanently. I am happy to run that test early in an engagement, including when the honest result is that a channel is not pulling its weight.

On LinkedIn, and when it is worth the premium

I am certified on LinkedIn Ads and I use the platform, so take this as a practitioner's view rather than a knock. LinkedIn charges a real premium per impression compared with reaching a similar person elsewhere. Sometimes that premium is obviously worth it: when your targeting genuinely depends on job title, seniority, company size or industry, and the deal size is large enough that a small number of right-fit conversations pays for the whole program.

Often, though, a team is paying that premium to reach an audience they could have assembled more cheaply somewhere else, because LinkedIn felt like the professional choice. If your buyer is reachable by interest, behavior or a customer-list match on a cheaper platform, and your deal size is modest, the math frequently does not hold. I will run the comparison before we commit budget rather than after.

Creative is the variable you still control

Targeting controls have narrowed across every major platform over the last several years, and the delivery systems have gotten better at finding the right person on their own. What you hand them to work with is now the main variable you control. Three near-identical ads give the delivery system nothing to choose between.

So the work is a steady supply of genuinely different concepts rather than color variants: new hooks, unfamiliar formats, arguments you have not made yet. I will tell you which concepts are worth producing more of, and I am comfortable saying that a piece of creative you paid a lot for is not working.

How engagements usually work

Monthly management, or a scoped project such as an account audit, a new channel launch or an incrementality test. If you are already running paid social and something feels off about the numbers, the audit is usually the right starting point, and it is the least expensive way to find out whether I can help.