A managed IT firm in Columbus puts $3,000 a month into LinkedIn Ads, aimed at operations managers at companies with 50 to 500 employees. Three weeks in, the dashboard looks better than anything they have seen on the platform. Click-through rate above 5 percent. Clicks under $2, on a platform where they were braced for $12. The owner forwards a screenshot to their business partner with one word: finally.
Then they open the website analytics, and the visits are not there. The contact form is quiet. The sales team has nothing new to call.
Nothing is broken. The campaign did exactly what LinkedIn set it up to do. Somewhere in the ad set settings sits a checkbox called LinkedIn Audience Network, it was checked when the ad set was created, and most of that budget never ran on LinkedIn at all. This is the most common surprise in a small LinkedIn account, and it is worth understanding properly, because the fix takes seconds and the cost of not knowing can run to most of your budget.
You Pay LinkedIn for Who People Are. Off the Platform, It Guesses.
LinkedIn Audience Network is LinkedIn's network of partner apps and websites. When it is switched on, your ad can appear outside LinkedIn entirely: as a native ad placed inside an article on a partner site, or as an interstitial, the full-screen kind that pops up between screens in a mobile app. LinkedIn says the same targeting, bid and budget you set for the feed carry over, and that the network "can help result in lower cost per key result."
It is switched on automatically for every new single image, carousel, document and video ad set. You do not opt in. You have to opt out, ad set by ad set.
Here is the part that matters for your money. On LinkedIn itself, targeting is the whole product. When you aim at operations managers at 50 to 500 person companies, LinkedIn knows who those people are because they told it, on a profile they keep current for their own careers. That knowledge is what you pay LinkedIn's premium for. Nowhere else online can you buy it.
Off LinkedIn, that certainty thins out. LinkedIn's own help pages explain how it decides who is who on a partner site: when a member is signed in, it matches them against your targeting. But "in some instances, we use a model to estimate whether the signal likely belongs to a member account, such as when a member account isn't logged into LinkedIn or is visiting a third-party site or app on an iOS mobile device." That covers a lot of people. Anyone browsing on an iPhone. Anyone who is not logged in at that moment.
So the expensive thing you came to LinkedIn for, knowing exactly who sees your ad, becomes an educated guess the moment the ad leaves LinkedIn. You are still paying LinkedIn's rate for the targeting. You are just no longer getting LinkedIn's version of it.
We Told You to Let Meta Pick Your Placements. LinkedIn Is the Exception.
If you read our Meta coverage, this may sound like a contradiction. In Meta Wants to Choose Where Your Ad Runs. Let It., we argued against unchecking Meta's placements, Meta's own Audience Network included. That advice still stands, for Meta. The reasons it works there are exactly the reasons it fails here.
On Meta, the targeting is behavioral everywhere. Meta does not know your buyer's job title on Instagram any better than it does in a partner app; it is predicting either way, and its system is built to buy the cheapest results across every placement for the budget you set. Taking placements away mostly forces it into its more expensive corners.
LinkedIn's value is different in kind. What you pay for is professional identity: title, seniority, company, company size. That identity is strongest inside LinkedIn, where people are signed in and acting as themselves at work. Push the ad outside, and you keep the price while losing much of what justified it. On Meta, more placements give the system more room to find buyers. On LinkedIn, the Audience Network mostly gives it more room to find cheap clicks.
Audience Network Clicks Cost Less Because They Are Worth Less
Every practitioner who audits LinkedIn accounts describes the same tell. AJ Wilcox, who runs an agency dedicated to LinkedIn Ads, says he can spot an account with the Audience Network left on immediately, because the top-line numbers look amazing: click-through rates of 5 percent or higher and clicks under $2. Then you look for the traffic and the conversions, and they are not there.
There are a few reasons the numbers flatter it.
LinkedIn adds a button you did not ask for. Audience Network placements require a call-to-action button with a link. If your ad does not have one, LinkedIn adds one automatically for those placements, and its help page says plainly that this "will impact your Clicks to Landing Page metric." More buttons, more clicks counted.
Apps produce accidental taps. A full-screen ad between levels of a mobile game gets tapped by thumbs trying to close it. Those taps count as clicks. Nobody reading a LinkedIn feed at their desk does that.
Some apps manufacture clicks. The same agency founder describes app developers who generate simulated clicks behind the scenes to collect ad revenue, and one client account where 60 percent of the budget went to a single app. LinkedIn is not naive about this. It runs automated fraud checks with partners like Pixalate, DoubleVerify and HUMAN on every bid request, and practitioners note that LinkedIn only charges for clicks it believes are human. But LinkedIn also states that impressions and clicks on the Audience Network are not accredited by the MRC, the industry body that audits how ad platforms count. On LinkedIn itself, you are trusting LinkedIn's counting. Off LinkedIn, you are trusting thousands of publishers' signals that nobody independent has validated.
It spends fast. Practitioners also report that the network does not pace spend the way LinkedIn's feed does, and can work through a full day's budget within about 20 minutes. That is a practitioner observation, not a LinkedIn statement, but it matches what many account owners see in their hourly delivery.
What does that look like in a real account? One practitioner, John Stewart, shared the On/Off Network breakdown from a retargeting campaign his agency ran on both. The Audience Network clicks cost about one-sixth of what the LinkedIn clicks cost. Its conversion rate was 0.3 percent, against 6.46 percent on LinkedIn. So the cost of each actual conversion on the Audience Network came out more than three times higher, despite the far cheaper clicks. That is one advertiser's account, not a benchmark, but the shape is the one everyone reports: cheaper clicks, far fewer results, more expensive results.
This is the same mistake as judging LinkedIn by its cost per click in the first place, which we unpacked in LinkedIn Isn't Failing. You're Measuring It Wrong. The click is an input. The meeting, the proposal and the signed contract are the outputs.
Campaign Manager Already Shows How Much of Your Budget Got Spent Off LinkedIn
You do not have to take anyone's word for this. Your account already holds the answer, and checking takes about two minutes.
Step one: the On/Off Network breakdown. In Campaign Manager, open the Campaigns, Ad sets or Ads tab and choose On/Off Network from the Breakdown menu. Each row splits into two lines: what ran on LinkedIn and what ran on the Audience Network. Look at spend first, then clicks, then conversions. If the Audience Network line carries a large share of spend and a sliver of conversions, you have your answer. To keep a copy, use Export and choose "Audience network ad performance."
Step two: the publisher reports. In the Brand Safety & Suitability Hub you can download the list of publishers in the network, and practitioners point to monthly publisher delivery reports going back about six months that show where your ads actually ran: the app or site, whether it was web or app, and its content category. Open one. Most owners recognize very few of the names.
Step three: your website analytics. In Google Analytics, traffic from the Audience Network tends to show up as referral traffic from the partner site, not as paid social from LinkedIn. Compare the two side by side: time on site, pages per visit, form fills. The difference is usually obvious.
Two quirks to know while you look. Some rows may be labeled Uncategorized, which LinkedIn says means the placement is unknown, often because the number was modeled rather than observed. And LinkedIn reports Audience Network conversions at the ad set level only, so you cannot drill into which individual ad converted off LinkedIn.
A Lead Campaign Has No Business Running on the Audience Network
Here is our position for a small B2B business: if an ad set exists to send people to your website to become leads, the Audience Network should be off. Not tested, not trimmed. Off.
The reasoning is simple arithmetic. A small LinkedIn budget buys a limited number of clicks from a limited number of real decision-makers. Every dollar the Audience Network takes is a dollar not spent putting your offer in front of an operations manager scrolling LinkedIn between meetings, the one setting where you can be sure who they are and that they are thinking about work. You cannot afford to trade those impressions for cheap taps in a game app.
To turn it off on an existing ad set, open the campaign, go to the Ad sets tab, hover over the ad set and click Edit. In the Placements section, click Change and uncheck LinkedIn Audience Network. On a new ad set, the same checkbox sits in the Placements section after you choose your ad format. (LinkedIn renamed its levels recently: what used to be a campaign group is now a campaign, and what used to be a campaign is now an ad set. The checkbox lives on the ad set.)
Two habits keep it off. First, LinkedIn's help pages describe the setting only ad set by ad set, and every new ad set starts with it checked, so make unchecking it part of your launch checklist. Second, duplicating an ad set copies its settings, so a clean original produces clean copies. One practitioner reports that choosing the lead generation objective, which uses LinkedIn's built-in forms, unchecks the box on its own. Check it anyway. It takes one second and LinkedIn's interface changes often.
This sits alongside the other defaults worth questioning on day one, which we covered in LinkedIn Ads Aren't Expensive. Your Setup Is., including the bidding default explained in LinkedIn Has Three Ways to Bid. It Defaults to the Priciest. And if you already turned off weekend delivery as described in Your LinkedIn Budget Spends All Weekend. Your Buyers Are Not There., LinkedIn will have made you turn the Audience Network off for that ad set already, because it cannot promise your schedule will be respected off LinkedIn.
The Audience Network Earns Its Keep in Two Places: Retargeting and Tiny Audiences
We would not tell you it is useless, because it is not. The fair case for the Audience Network is about reach. Not everyone in your audience opens LinkedIn every day. Some of your best prospects check in once a month. The network lets you stay in front of those people while they read the news or check the weather.
That reach is worth paying for in two situations.
Retargeting people who already know you. Someone who visited your pricing page, watched most of your video or opened your lead form has already shown interest. Reaching them again off LinkedIn is a reminder, not an introduction, and the identity question matters less because they came from you in the first place. Practitioners who defend the network almost always defend it here. If your warm audience is still small, You Don't Have the Website Traffic to Retarget. LinkedIn Will Retarget Anyway. covers how to build one without much site traffic.
A small audience you have already saturated on LinkedIn. If you sell to a narrow group, say the facilities directors at a few hundred hospitals, you can reach most of them on LinkedIn within weeks and start showing the same people the same ad too often. At that point the Audience Network is one way to find extra impressions with the same people elsewhere, as an alternative to pouring more budget into a ceiling, the problem we described in More LinkedIn Budget Buys the Same 50,000 People Again.
Notice what is missing from that list: cold prospecting to strangers. That is the job LinkedIn's verified targeting exists to do, and it is precisely the job the Audience Network does worst.
An Audience Network Test Gets Its Own Ad Set and a Tenth of the Budget
If one of those two cases fits and you want to test it, set the test up so it cannot quietly take over the account. The practitioners who use the network well agree on almost every rule.
Split it out. Duplicate your existing ad set. Leave the Audience Network off on the original and give it about 90 percent of the budget. Turn the network on in the copy and give it about 10 percent. Even with the box checked, an ad set can still run in the LinkedIn feed, but the network tends to win the budget, so the split is what keeps it honest. Keeping the test in its own ad set also keeps its unusually cheap clicks from blending into your real numbers.
Use an allow list, not just a block list. One practitioner counted more than 500 publishers in the network and recognized almost none of them. Blocking the bad ones one at a time is a losing game. Instead, download the publisher list, pick the few dozen sites your buyers actually read, and upload those as an allow list so the ads run only there. You can apply a list to a single ad set or to every active ad set in the account through the Brand Safety & Suitability Hub.
Exclude whole categories. LinkedIn lets you exclude publishers by content category. Mobile games, uncategorized sites and other categories your buyers would never associate with work are the usual first cuts. The categories come from LinkedIn's partner Pixalate and can change during a run, so LinkedIn itself recommends reviewing them as you go.
Bid manually and cap it. Do not let the test ad set run on automatic bidding. A manual bid keeps the network from chasing volume.
Judge it in your CRM, not in Campaign Manager. A higher click-through rate is not a result. After a month, compare the two ad sets on what reached your sales pipeline: qualified leads, meetings booked, proposals sent. If the Audience Network version cannot show you real pipeline at a cost you would happily pay, turn it off and move the 10 percent back.
LinkedIn Checked the Audience Network Box for You and Never Mentioned It
None of this requires a bigger budget or a new strategy. It requires five minutes this week. Open Campaign Manager, run the On/Off Network breakdown for the last 90 days and see how much of your spend went off LinkedIn and what it produced. Uncheck the box on every ad set that exists to generate leads. Add "Audience Network off" to the checklist you use before any new ad set goes live. And if you have a real retargeting or saturation case, test it in its own fenced-in ad set and let your pipeline decide.
For most small B2B advertisers, that single checkbox is the difference between paying LinkedIn prices for LinkedIn buyers and paying LinkedIn prices for game-app taps. If you would rather have someone go through your account and set these defaults right, that is the kind of thing our LinkedIn ads team does every day. If you would rather do it yourself, you now know exactly where to look.
