Where B2B Ad Budgets on LinkedIn Actually Burn

When your B2B LinkedIn ad campaigns underperform, everyone looks at the creative: new visuals, new copy, new formats. The audience definition, the thing that decides who sees any of it, gets a fraction of the attention.

That’s backwards. Based on years of running LinkedIn ads, from global programs to tightly scoped local ones, I can tell you where the money actually leaks. It’s rarely the ad. It’s the audience.

There are four common leaks.

Leak 1: Your targeted audience is far too wide

LinkedIn recommends ad audience sizes starting from around 50,000 people. Follow that advice for a B2B ad campaign and you’ve already lost.

How many companies genuinely fit your ICP, and how many relevant decision-makers work there? For most B2B businesses, nowhere near 50,000. The rest is padding: people who look vaguely relevant but will never buy from you.

In our experience, smaller audiences consistently perform better, sometimes down to a few hundred people, as long as the platform still lets the campaign run. Your reach numbers will look modest. That’s fine. You’re not buying reach, you’re buying relevance.

Leak 2: Your targeting parameters are too thin

The second leak is stopping at the basics: a few job titles, an industry, a geography, done.

To actually land inside your ICP, go one level deeper. Either target at the company level with an uploaded list (or CRM integration), or define industry and geography tightly enough that the combination does the work. Use either a well-established job title grouping or work with seniority and job functions.

The catch with industry targeting: companies choose their own industry labels on LinkedIn, and they choose them inconsistently. You’ll miss relevant companies that tagged themselves differently and catch irrelevant ones that picked your tag. It looks precise on screen. In reality, it’s fuzzy.

The good news: with AI deep research tools, building and verifying a target company list takes a fraction of the time it used to. There’s little excuse left for skipping it.

Leak 3: You set the audience and walked away

Even with strict job titles and company lists, even targeting a single company, your ads will not stay neatly inside the lines. We’ve seen it repeatedly: tightly targeted campaigns still leak impressions outside the intended audience. The platform’s matching errs on the side of showing your ads to more people, not fewer.

The fix is unglamorous: monitor where your impressions actually land, and build block lists. Yes, even for companies you never targeted in the first place. It feels absurd to block companies you didn’t invite, but that’s the reality of the platform.

Leak 4: You trusted the “expanded audience” checkbox

LinkedIn’s audience expansion and lookalike features promise more scale with one click. In B2B, they’re usually a bad idea. Test them if you must, carefully and in isolation, but know that the default outcome is diluted targeting and diluted data. The burden of proof is on the feature, not on you.

A word on positive leaks

A smart man once told me about positive leaks: things that are going well, but could be going better. Even an ad campaign that hits its targets usually hides a few of them. So even if everything looks fine, it might bear a check, and that’s a check we’re happy to run for you.

Why this matters more than the money

The obvious cost is monetary. You think you’re running targeted ads, but really they’re semi-targeted. Targeted in your head, scattered in reality.

The less obvious cost is worse: polluted learnings. Every campaign is also a test. If those tests run against the wrong audience, every signal you collect is contaminated, and you’ll carry the false learnings into the next campaign, and the one after that.

This also settles the testing order. In a new audience, test the audience first, then the creative. We simply don’t care how well creative performs in front of the wrong people. The opinions of non-ICP are worth exactly nothing to your business, no matter how many of them click.

Remember who you’re dealing with

LinkedIn is a shop, and you’re standing at the till. Every recommendation inside the platform, the audience size suggestions, the expansion checkboxes, the optimization prompts, should be read the way you read “would you like fries with that?” It’s added sales. LinkedIn is in the money-making business, and it’s very good at it.

That doesn’t make the platform bad. It makes it a platform. Use it on your terms, not its defaults.

Set it and forget it is how you lose

We’re a bit lazy by nature. We want to build the campaign, hit launch, and check back at the end of the month. Such are people, such is you, such is me.

Unfortunately, that’s not the way to victory. Audience hygiene is daily-to-weekly work: watching where impressions land, tightening parameters, updating block lists, and adjusting as new ad tests roll out. Skip it, and your quarterly review becomes an archaeology dig into where the money went.

Also read:

5 Things You Need to Know About LinkedIn Marketing in 2026
LinkedIn Thought Leadership Advertising Utilizing Executive’s Personal Brand
5 Examples of LinkedIn Advertising Strategies for B2B

Growth Marketing Team now available with predictable budget

See pricing