Why LinkedIn ads look like they fail in Google Analytics for B2B SaaS
Your LinkedIn ads are running. The CFO opens Google Analytics, filters by channel, and finds LinkedIn near the bottom. A handful of conversions. A cost per acquisition that looks embarrassing next to Google Search.
So the question lands on your desk. Why are we still paying for LinkedIn?
Sound familiar?
This is one of the most common conversations we have with B2B SaaS and IT companies. The ads are working. The analytics say they are not. And the gap between those two things is quietly killing budgets that should have stayed.
Last-click was built for a buying journey B2B does not have
Google Analytics, by default, gives all the credit to the last click before a conversion. For an ecommerce store selling a €40 product, that is roughly fine. Someone clicks an ad, buys, done.
B2B SaaS does not work like that. A CTO sees your ad on Monday. She does not click. She reads a post from your founder two weeks later. She still does not click. A month on, she remembers your name, types it into Google, lands on your site, and books a demo. Google Search gets the credit. LinkedIn gets nothing.
The ad did the work. The search box took the applause.
Why this hits IT and SaaS harder than most
The longer and more considered the purchase, the wider the attribution gap. A high-value B2B contract involves multiple people, several months, and a lot of touches that never end in a trackable click.
LinkedIn's strength is exactly the part analytics cannot see. It puts you in front of the right decision maker before they are actively searching. It builds the familiarity that makes them click your search result instead of a competitor's. That influence is real, but in a last-click report it is invisible. So the channel that warmed the lead gets blamed for not closing it.
The three places the real signal actually lives
If the dashboard lies, where do you look instead? Three places.
First, ask your leads directly. A single "How did you first hear about us?" field on your demo form will surface LinkedIn far more often than analytics ever does. It is not perfect, but it captures the memory that tracking loses.
Second, watch what happens to your branded search and direct traffic when LinkedIn ads run. Turn a campaign on for a strong region and you will often see people searching your name and typing your URL go up a few weeks later. That lift is LinkedIn working, showing up in a channel that takes the credit.
Third, look inside your CRM, not your analytics. Tag opportunities by first known touch and compare deal size and win rate. In a lot of B2B pipelines, the deals that had an early LinkedIn touch are larger and close more reliably, even when the last click came from somewhere else.
What to change before you cut the budget
Do not judge LinkedIn on a metric it was never going to win. That does not mean flying blind. It means measuring the channel on its actual job.
Set up a view that combines self-reported source, branded-search lift, and CRM-level first-touch. Give a campaign a fair window, several weeks, not several days, before you decide. And separate the two questions that usually get mixed together: is the campaign generating demand, and is your site converting that demand? A LinkedIn campaign can be doing its job perfectly while a weak landing page quietly wastes it.
Cutting a channel because of a last-click report is how good pipelines get starved. The budget looks like a loss on one screen and a driver of growth on another. The screen you trust decides what you do next.
Where TriAds fits in
We run LinkedIn ads for B2B SaaS and IT companies that take their budgets seriously, from €2,500 a month up. Most of our clients spend more. We build the ICP, run the campaigns, and give clients a live dashboard at my.triads.marketing, so what you see is not a last-click summary weeks late but what is happening now, tied back to pipeline instead of just clicks.
If your analytics say LinkedIn is failing but your best deals keep mentioning it, the measurement is usually where the answer lives.