©2026 ZenABM - All Rights Reserved.
LinkedIn ads mistakes can burn through your budget faster than on any other ad platform (the CPCs see to that). Most of them don't even look like mistakes: CTR is fine, CPC is fine, the dashboard is green, and there is zero pipeline. You can make them at every stage, from strategy and audience to the landing page and the reporting, so I've sorted all 57 by the stage where you make each one. This is a skill issue, not a platform problem, and every one of them is fixable.
Sources: Max Herczeg's session at the ZenABM ABM Bootcamp (2.5 years inside LinkedIn), the 2026 benchmark report (211 companies, 161,256 ads, $5.5M spend), 516 posts from practitioners who publish their real numbers, and my own scar collection.
/linkedin-abm-audit skill returns a fix list ordered by dollars at stake, and multi-channel attribution shows what touched each deal. Plans start at $59/month billed annually.Three numbers to keep in your head while you read the detail:
Every mistake in this post is below with a one-line reason, so you can tick them off against a live account. Anything you can't tick is costing you money right now. I turned my own list of LinkedIn ads mistakes into a Halloween costume joke on LinkedIn, which tells you how often I see the same ones:
These are the expensive ones, because everything downstream inherits the problem, and I made most of them. If you want the full program story behind them, I wrote it up in running ABM on LinkedIn.
Max Herczeg is an independent LinkedIn ads consultant who spent 2.5 years at LinkedIn. He taught the session Campaign Setup & Avoiding Mistakes That Kill ABM Performance at the ZenABM ABM Bootcamp, our webinar series on running ABM on LinkedIn. He says planning is what most new clients skip:
"Most of the times, when I get with clients, they want to start with targeting and the campaigns right away. But they forget the non-negotiables and the planning." Maximilian Herczeg
Here is the strategy framework Max starts every client engagement from:
You want $200k in pipeline a month and you're spending $5k on ads, but how did you arrive at either number? I didn't know for a long time. To calculate your realistic budget, work backwards:
A quick sanity check comes from the 2026 benchmark report: the median is $5.21 of influenced pipeline per $1 of ad spend, and top performers reach $15.20. At $5k a month that is about $26,050 of influenced pipeline at the median, or $76,000 at the top, so a $200k a month goal on $5k was never realistic.
You can also run it Kirill Vdov's way: ICP size / 2 / 1000 x CPM x 6 / 0.75. The "/ 2" matters because only about half of any LinkedIn audience is reachable in a month. As Kirill, a LinkedIn ads consultant, put it:
"Yes, it costs that much to saturate just 50k people with an average LinkedIn CPM." Kirill Vdov, on LinkedIn
Then cross-check against the account funnel numbers I learned the hard way: about 5% of your target accounts are in market, you'll reach 50% to 70% of them, and about 0.45% will get to your landing pages.
Want to skip the manual calculation? Use the free ABM Budget Calculator, which plans your LinkedIn ABM budget from your targets and account list. If you're on ZenABM, the same math runs as the /abm-strategy-planning skill on the ZenABM MCP server (connect Claude to https://app.zenabm.com/api/mcp, then run /init):
Run /abm-strategy-planning for my account. Use my actual CPC and CTR by ad format from the last 90 days, plus ACV and close rate from the CRM, against a pipeline goal of $200k a month. Tell me the audience size and monthly budget I need, and whether my current target list is too big for what I'm spending.
Audience penetration is the number that tells you whether your budget and audience even match. Here are the audience penetration benchmarks I use:
Patrick Cumming, a paid media consultant, puts the problem simply:
"Targeting a 500K audience with $5K a month won't work. You'll get 5% reach and 2 frequency." Patrick Cumming, on LinkedIn
These are the frequency targets to size against:
Metric | Cold audiences | Warm audiences |
|---|---|---|
Monthly audience penetration | 30% to 50% | 70% to 90% |
Impressions per person per month | 6 to 10 in total across campaigns | 10+ |
Impressions per creative before fatigue | 2.5 to 3 per person (TLAs start fading from about 2) | 2.5 to 3 per person (TLAs from about 2) |
Marc Woodland's version after 150+ audits is a simple question: do you want to reach 500,000 people once, or 50,000 people ten times? For ABM it's never the first. Marc also has a Rapid Fire episode on audience sizing in our show if you want the long version.
LinkedIn needs at least 300 members to run an audience at all, and it recommends 50k. I'd ignore the 50k, because it exists for LinkedIn's benefit. Kirill runs ABM audiences of 5k to 15k and segments anything over 100k.
I made the opposite mistake for ages, slicing campaigns thinner and thinner for "precision". The thinner you slice, the more expensive your campaigns and the thinner your ROAS. I now run 1 cold plus 1 warm campaign and max 2 personas, and it beats everything more elaborate I tried.
"Most teams do it in reverse and then blame the creative." Justin Rowe, founder at Impactable, on LinkedIn
Across his 200+ accounts, warm first-party audiences (CRM, site visitors, video viewers) look very different from cold ones:
Cold broad targeting | Warm first-party audiences | |
|---|---|---|
Cost per lead | $300 to $600+ | $75 to $250 |
Close rate | Baseline | 3 to 5x better |
Leads on $10k a month | About 22 (all cold at $450 CPL) | About 49 (half moved to warm) |
Don't seed your target list with cold lookalikes either. Start with closed lost and churned accounts, because they already know you, they've already evaluated you, and the cycle is shorter.
Ishaant Shakunt from SpareGrowth ran the "Book a Demo" CTA through his offer scoring framework at the Bootcamp, and only 2 out of 10 checks passed. His verdict: "This is so bad that it's not worth tweaking it. Just don't run it."
For the receipts, Kristof Bardos documented $8k spent, 74 clicks and 0 leads from cold traffic pointed at a form. Offer something worth the trade instead (audits, calculators, benchmarks, templates), and test at least 4 offers, because the one you're sure will win usually doesn't.
Ivan Falco, an ABM consultant, explains why a lead-gen-only budget stalls:
"At any moment, only ~3% of your market is in-market. Lead gen fights over that small pool. Drain it, and next month you're back fighting again." Ivan Falco, on LinkedIn
You can watch the cost curve prove it. One company on a $50k a month bottom of funnel budget paid $200 to $300 per demo request, then $600 to $800 within weeks, then $1,200+, on a $40k to $60k ACV, with ROI still negative. Max's version at the Bootcamp was that 100% in lead gen stagnates fast and lead prices skyrocket.
Kirill's rule is to run for half your sales cycle, minimum 3 months, because 3 months is what it takes to reach everyone even once. Judge on SQLs and opportunities at that point, not closed won, which lands long after your kill decision.
The classic counter-example is a cold 84k audience with no warm-up, run for 4 days on $2k. It got 43 clicks and 0 leads, and the verdict was "LinkedIn doesn't work".
Here is the ICP planning slide on segmenting only when personas need different messages:
This is where most of your budget dies, and you can't see it in the dashboard, because the wrong people click at the same rate as the right ones. Max's rule:
"If you mess up the audience, everything else is really pointless. This is not Meta, this is LinkedIn. The algorithm is not gonna find the right people for you." Maximilian Herczeg
The Audience Hub shows who you're about to pay to reach before you spend anything. Here is the Audience Hub interface:
"Use the Audience Hub. It's the most important thing when it comes to campaign setup." Maximilian Herczeg
Go to Plan > Audiences > Saved > Create Audience. Here is Max's 5-minute pre-launch check:
Re-check after every exclusion change. I also found a bug the hard way: edits to a saved audience don't propagate to the ad sets already using it until you re-select the audience in each one.
When you type in a job title, LinkedIn expands it into a "super title", a bundle of titles it considers similar, and the bundles get wide fast. Here is the job title targeting data that shows the expansion:
"If you put in VP of Engineering, you'll also get engineering specialists, infotech specialists, construction specialists." Maximilian Herczeg
His example was a 20,000 person marketing-leadership audience sitting inside a "marketing specialist" super title of 100,000+. You also end up bidding in the super title's auction, not yours, which is why your CPMs come in higher than your targeting suggests.
The fix is to check one title at a time in the Audience Hub, exclude the junk it dragged in, and switch hopeless personas to job function plus seniority. Job title engagement in ZenABM groups ad engagement by job title, so you can check whether your spend actually reached the buying committee.
"For a list to make sense on LinkedIn, you need at least the LinkedIn URL of a company page. We're aiming for a match rate of 90% and more." Maximilian Herczeg
Without the URL, LinkedIn guesses. Ivars Krutainis collected the guesses: No Cow organics matched to Cow PR agency, Pendulum probiotics to an architecture firm, and Beam organics to an IT company. Enrich your list with LinkedIn URLs first (Clay does this well), upload that column, then spot check by hand. Here is the company list upload dialog:
This is my favourite scar, because it undoes the previous mistake without telling you. LinkedIn told me 90%. I downloaded the audience it actually built and compared it with my list, and the real match was 43%, full of false positives from similar company names.
A separate HubSpot-pushed audience reported 90%+ and contained a third of my companies, plus the South African Post Office. Getting to a real 87% took my ABM demos up nearly 100% month over month.
The fix takes four steps:
Then verify in the wild. Company engagements in ZenABM names every company that saw or clicked each campaign, mapped against your target list:
Companies self-select their industry, usually years ago, usually via someone who's long gone, so Amazon can sit under Internet, Retail or Cloud Computing. Here is Amazon's LinkedIn page with its industry tag:
Kristof Bardos measured the damage. An account targeting founders of 2 to 10 person software companies spent $13,852 for 0 leads. Six of the top 10 engaged companies were HR firms, agencies, hospitals and restaurants, and $8,034 (58% of budget) went outside the ICP. For ABM, upload your list instead.
Exclusions always beat inclusions. If someone has two active roles, one at a company you target and one at a company you exclude, that person is gone. That happens constantly with advisors, board members and side businesses. Here are the exclusion filters in the targeting interface:
"I would not recommend excluding seniorities. Just exclude job titles." Maximilian Herczeg
LinkedIn's seniority data is unreliable: Max showed engineers with 12+ years of experience classified as "entry". If you exclude "entry" to remove interns, you remove experienced ICs with them. So exclude specific job titles only, never seniority or function.
The opposite failure is just as common, which is no exclusions at all. Remove students, competitors, job seekers and customers from prospecting, but don't exclude customers from everything, or the only vendor ads they'll see are your competitors'. ZenABM's Exclusions apply to companies, job titles and industries team-wide or per campaign, including campaigns you create later.
"LinkedIn defaults to showing ads to people who've visited a location. Change to permanent." Marc Woodland, LinkedIn ads consultant, on LinkedIn
The default ("Recent or Permanent") includes anyone who visited the country in the last 6 months, so you're paying to advertise to tourists. In one account Marc audited, this plus no size filter produced a 1.3M audience, $800 spent and 1 lead, and seven minutes of fixes took it to 54k. Changing the setting takes one second.
Most accounts never build these, and they're your cheapest audiences:
Put ICP filters on all of them, because a visitor audience with no filters is just traffic.
I use LinkedIn's current names here: a campaign (formerly campaign group) holds ad sets (formerly campaigns), and ad sets hold ads. Structure decides where your money lands before any setting gets a vote.
LinkedIn spends where impressions are cheapest and most available, so the biggest companies on your list eat the budget. In one Series D account at $45k a month, only 15% of target accounts got 3+ impressions in 30 days. Tim Davidson, founder at B2B Rizz, calls it out:
"most people just put all accounts into one bucket and let run LinkedIn ads do their thing. BIG MISTAKE" Tim Davidson, on LinkedIn
His fix is a waterfall, with nothing else changed (same budget, targeting, objectives and bidding):
One more split to make is US and Europe. LinkedIn doesn't balance your daily budget across time zones, so the US spends it before Europe wakes up.
With 15 ad sets at $500 a month each, none of them hit frequency, and LinkedIn needs a minimum spend per ad set to optimise delivery at all. Run $50 to $100 a day per ad set with 4 to 6 ads each, and if your budget is tight, consolidate.
Ivars Krutainis found one campaign with 52 live image creatives, some with zero clicks. 20 ads on $100 a day is $5 per ad per day, so nothing exits learning and you learn nothing.
How many ads can you run? Use monthly budget / 30 / (average CPC x 3) = number of ads, which gives each ad about 3 clicks a day. Take $3,000 a month: that's $100 a day, and at an $11 CPC (an example, use yours) 100 / 33 is about 3 ads. The free LinkedIn Ads Count Calculator does it for you.
Most accounts run all single image ads, all pointed at conversion. Each source below measures something different, so compare within a row, not across rows:
Source and metric | Thought Leader Ads | Single image | Company page content |
|---|---|---|---|
ZenABM benchmark report, landing page click cost | $3.06 (77% cheaper) | about 4.3x more expensive | Not measured |
My own account, TLAs with a link: landing page CTR / landing page CPC | 0.90% / $12.73 | 0.43% / $13.71 | Not measured |
Gabriel Ehrlich, 30+ accounts: engagement rate / cost per engagement | 10% to 18% / under $1 | Not measured | 0.5% to 1% / $8 to $15 |
The numbers move when budget moves to TLAs. I shifted TLA share from 41.2% to 52.8% (cutting single image) and got roughly 2x the demos on the same budget. Gabriel's client also took cost per SQL from $4,270 to $923 in one month by moving budget to leadership posts.
I'd also add a warning on document ads, because I don't use them. ZenABM's own document ads underdelivered at 6,913 against a 15,700 impressions target, ran at a $42 CPM against a $26 target, and cost $12 to $13 per landing page click against $11. A report download is also very low intent, and document ads don't let anyone book a demo.
Text and Spotlight ads sit in the right rail and charge per click or per impression, and nobody clicks them. Valentin Dimitrov runs them at about $0 spent for 10k to 100k impressions.
The funnel looks great in a deck, but you have no idea where anyone is in their buying process unless they tell you. Max says there's no step-by-step process for getting cold people to convert, and people can very much convert in a cold audience on day one. Tim Davidson runs cold plus warm instead (without excluding warm from cold), and for some companies 80% to 90% of demos come from the cold audience.
You can build a pristine list in Clay, debate ICP for 3 hours, and then hand it all back to LinkedIn's defaults. The defaults are built for LinkedIn's revenue, not yours, and the most expensive settings are the ones nobody opened.
All four are on by default in a new campaign, and I'd turn all four off every time. I think the hiding is deliberate, as I said in our Rapid Fire LinkedIn Ads x ABM Show episode with Valentin Dimitrov, and even experts like Tim Davidson sometimes leave a harmful setting on. Here are the budget and bidding settings in Campaign Manager:
To check your own LAN damage, open your campaign, then Breakdown > On/Off Network. Brand Safety > Publisher Delivery Reports shows the actual sites.
I also create campaigns in the classic editor, not Accelerate, as I explained in our LinkedIn Ad Campaign Setup Review & Launch webinar. Accelerate and LinkedIn's AI settings push you into settings that may not be good for your spend.
ZenABM's Budget Savers switches off Audience Network and Audience Expansion across all ad sets (or a chosen scope) and re-applies the setting every hour, so a new ad set can't ship with either one on.
Manual bidding hides behind "Show additional options", and Max has a story about it:
"Manual bidding is so good that LinkedIn actually hides it from you. Maximum delivery, cost cap, and then manual bidding is underneath, which doesn't make any sense." Maximilian Herczeg
Here is Max's manual bidding process, shown in the flowchart below:
His one exception for Maximum Delivery is TLAs on a Brand Awareness objective, where he sees higher reach and lower CPMs. It's also fine when you're out of time (webinar in 2 weeks), as long as you watch the CPM. ZenABM's Bidding agent checks every bid change against LinkedIn's suggested bid before you commit, and changes budgets in bulk across ad sets.
Here is the counterweight to Max's process. You pay one cent above the second-highest bidder, so a low bid doesn't make clicks cheaper, it caps your reach. In my setup webinar I also say not to open above the recommended bid, because you'll overpay and push up the whole auction before you know whether the campaign is delivering.
"Manual bidding ruins 17% of my campaigns." Kirill Vdov, on LinkedIn
A low bid oversaturates the cheap 30% of your audience and spikes frequency there, so the expensive majority never sees you. Kirill's fix is to raise the bid 30% to 80% if results are bad. The failure is delayed, because everything looks great until the cheap segment runs out.
The objective decides what LinkedIn optimises toward. Here is the objective selection UI:
"Don't use website conversions. It's actually not working well. It's a more expensive version of website visits." Maximilian Herczeg
Tim Davidson tested it (big audience, pre-seeded conversion, max delivery) and got $88 per conversion, versus $21.51 per landing page click on Website Visits (different metrics). Use this table as your reference:
What you want | Objective | Note |
|---|---|---|
Traffic or demo signups | Website Visits | Not Website Conversions |
In-feed consumption, retargeting pools | Engagement | You pay for social clicks too |
Video metrics and retargeting | Video Views | Optimise for impressions, not clicks |
Gated content | Lead Generation | Bid for clicks, they're cheaper than leads |
TLAs for reach | Brand Awareness | The only case Max uses it |
Confession: I spent ages telling people Brand Awareness was a waste, then had to apologise to it 😅. Tested on the same TLAs, Engagement won on CTR and dwell time. But Brand Awareness delivered 40,047 impressions against 10,075, 5.50% penetration against 1.80%, and about 10x lower cost per landing page click. They do different jobs, so test both.
Put Salesforce on a list with 500 smaller companies and Salesforce eats the budget. Kirill's number is that 2 or 3 large companies take 40% of impressions, and they're never in market. Here is the impression caps explanation slide:
LinkedIn has no native per-company cap, so you build one:
If you don't want to eyeball the threshold, use median ICP company size / 2 x 1.5 = your 7-day cap. Delivery and the Companies tab don't sync instantly, so companies can overshoot, but the cap still drags the worst offenders from about 40% of impressions to about 10%. Doing this by hand got old for me, which is why ZenABM has impression capping: it excludes companies past a threshold from some or all campaigns.
"Optimize for Performance" favours the highest-CTR creative, and Max has seen lead gen campaigns where the lower-CTR image had a much better cost per lead. Worse, your A/B test was never a test, because LinkedIn picks a favourite in the first days and starves the rest. Here are the rotation options in Campaign Manager:
"if you haven't switched on 'Rotate Ads Evenly,' you haven't run a test at all." Jack Cleere, paid media consultant, on LinkedIn
Worried that even rotation wastes money on losers? Tim Davidson tested it: after $12k, even rotation had a higher CPM but also a higher CTR, so CPCs came out nearly identical. Rotate Evenly at launch, then switch to Optimize for Performance to scale winners, checked against your metric, not CTR.
Max's rule for creative is to be bold, mix formats and react to the data. I'd add that no creative choice rescues a wrong audience, so fix Stages 2 to 4 first.
A 2019 WHM survey found 48% of B2B buyers find B2B ads boring and 82% want more creativity. In my analysis of 2,828 ads, stock photos appeared in 35% of the worst performers, 12% of the top ones and in none of the 17 highest-CTR ads. Use real people instead: your team, your customers, your founders.
TLAs work because they don't look like ads. Then people open with "we're excited to announce we've been featured in the Gartner report", write in third person and pitch the product, and get company-page numbers.
What works is a story or humour hook in the first 2 lines, real companies with real numbers, first person, admitted mistakes and value before any CTA. My best TLA campaign ran at $0.15 per landing page click (66x cheaper than image ads) on $391 total spend.
Put the link in the post body too. When I split mine out, TLAs with links averaged 0.90% landing page CTR at $12.73 landing page CPC against image ads at 0.43% and $13.71, and six link-less TLAs were dragging the format's average down. Links don't hurt paid reach the way they hurt organic.
CTR tells you the hook worked (and a curiosity trick can fake it), while dwell time tells you the content worked. Valentin Dimitrov, a paid social consultant, explains why B2B committees make this matter:
"You cannot fake someone sitting on your post for 12 seconds, reading the whole argument, and moving on without clicking anything." Valentin Dimitrov, on LinkedIn
His rules come from about $200k of TLA spend: a CTR above 10% is a strong TLA, and dwell above 10 seconds held for 4 weeks earns a permanent slot. For video, use completion rate. ZenABM puts dwell time and video funnel analytics next to job title analytics, so you see who did the staying.
Judge each format against its own CTR benchmark, since Campaign Manager's CTR counts every click:
LinkedIn audiences are small, so fatigue hits fast, and TLAs decay fastest (they start dying after a frequency of about 2). You can get it wrong in two ways: killing an ad on day 2 (a few hundred impressions tell you nothing, so wait 7 days) or waiting for CPL to climb, which is weeks too late.
"waiting for performance to drop before you swap creative means you're already behind competitors" Liam Collins, paid media consultant, on LinkedIn
Rotate when penetration and frequency plateau, and batch refreshes every 4 to 6 weeks with the next batch already in production. For a quick diagnosis, CPM up plus CTR flat means competition, and CTR down plus CPM flat means fatigue. ZenABM's Bidding agent flags ads whose click-through rate has been sliding and pauses them once you approve, so you don't track decay in a spreadsheet. Or ask Claude through the MCP server:
Compare CTR, engagement rate and dwell time week over week for every active ad over the last 8 weeks. List the ads where CTR declined three weeks running while CPM stayed flat (fatiguing, not facing competition). Separate critical cases from a watch list, and tell me which format is decaying fastest.
This is where a perfectly targeted, well-bid campaign goes to waste after the click.
Kristof Bardos documented the purest version: cold traffic to a Google Form, no warm-up, "here's our tool, fill in your details". It produced $8k spent, 74 clicks and 0 leads, which is a sequence problem, not a LinkedIn problem. High-friction offers fail the same way, and a webinar invite to a cold audience got 500 impressions and 1 click.
The ad promises a problem, the page talks features, you lose them at the click, and the campaign gets blamed. Tas Bober, founder of The Scroll Lab, taught a framework at the Bootcamp: build the page for where the person is, not who they are. A cold TLA click and someone who's read 3 case studies need different pages, and the page should read like a business case, because whoever clicks has to sell it to a committee that never saw your ad.
Here are examples of landing page personalization mistakes:
The cheapest fix before rebuilding anything is to make the page headline match the language of your best performing ad. For more, read ABM landing page examples that actually convert.
"when performance falls off a cliff out of nowhere, the ad is the real culprit less than 5% of the time." Liam Collins, on LinkedIn
Numbers dip, everyone swaps the hook and the image, and meanwhile it was a tracking break or a page change. Diagnose in this order every time: tracking, then page, then creative. Keep a change log and alert on a 15% week-over-week drop at any funnel stage.
Everything above can be right and the program still returns nothing, because the value gets extracted here or not at all.
A converted lead has a shelf life measured in minutes. Valentin Dimitrov saw two accounts lose in one month with everything else right: one sales team called back days later (the lead had booked a competitor), and the other had an alert system that never fired and nobody noticed for weeks. A 9% booking rate that should be 40% to 50% is a follow-up problem, so audit the first hour before touching any campaign setting.
Without a score, nothing triggers outreach and sales prospects blind. Justin Rowe's starting weights are: TLA click +1, company page visit +2, 30+ impressions +2, 75% video view +3, website visit +3, form submit +4, and 2 to 3 people from one company engaging +5.
Honestly, the scoring isn't the hard part. The hard part is defining "showing interest" in a way that actually correlates with buying intent, and most ABM programs break exactly there.
Account scoring in ZenABM scores every account from its ad engagement and CRM data and keeps the score current, with a recent and an all-time view, so "heating up now" and "warm for a year" stop looking the same. Customisable ABM stages sit on top, with your own names and thresholds. If you have no tooling at all, export your 7-day engaged companies list and hand the top 10 to sales on Monday, which takes 10 minutes.
"Man... if I showed this to our CFO, he'd stop the program today." Gonçalo Prates, ABM consultant, on LinkedIn
He checked impressions against the accounts sales was actively chasing and found 20, maybe 50 impressions company-wide in 90 days. Each month, take the accounts sales expects to talk to in the next 3 to 6 months and check paid impressions against that list, before your CFO does.
Engagement data is the raw material for your next conversation, not a reporting line. Outreach that opens with "hi, we're ZenABM" is weak next to outreach that opens with the topic that account has been reading about for 3 weeks. Here is a ZenABM contact list with company and intent data:
The play is to tag campaigns by feature, use case or offer, so engaging accounts inherit the intent theme. Push it to the CRM with CRM sync (HubSpot, Salesforce or Attio), or send engaged accounts to Clay or your sales tool over webhooks (Growth plan and up), and the rep opens with the topic. It's first-party data, built from engagement with your own ads.
I'd take it further. Out of 295 ABM-influenced deals in one 90-day window, 37 also had Google Ads touchpoints, 34 organic, 68 direct and 2 AI referrals, and no attribution model untangles that. So I use recent touchpoint combos as time-sensitive intent signals. If an account engaged with my product-analytics LinkedIn ad AND a competitor-alternative Google ad in the last 7 days, that tells me more than any score I ever built. Or ask the MCP server:
List every company that engaged with my ads in the last 14 days, has two or more people engaging, and has no open deal in the CRM. For each one show the intent themes they engaged with, their current and total engagement scores, and the last three campaigns they touched. Rank by current score.
On an expensive product, a whitepaper download is a signal that someone is curious, not that they intend to buy. It calls for nurture and social selling, not a cold call. Ask for the LinkedIn profile on the form, so social selling is even possible.
These don't waste budget directly. They make you switch off the campaigns that were working, which is worse.
The finding that shocked me most in my own data is that across the 33 companies with complete pipeline records, CTR correlates negatively with pipeline (Spearman rho -0.170). Spend correlates at 0.566 (and 0.511 across all 211 companies).
Once you see it, the trap is everywhere. Ad rotation optimises toward CTR, reports show one CTR column for every format, and the buyers your awareness campaigns exist for don't click at all, so a click-based cull kills the ads doing the heavy lifting. CTR is a creative diagnostic, not a success metric. Here is the influenced revenue and pipeline breakdown I'd report instead:
Cristina Maliţa, a LinkedIn ads consultant, says it bluntly:
"Don't run LinkedIn ads if you're going to judge their success on last-click attribution only. Seriously, don't do it." Cristina Maliţa, on LinkedIn
Her traced journey: 1,000+ LinkedIn impressions and clicks in January, an ebook in March, and a brand search in April, logged in the CRM as Direct Traffic.
Mine was worse. A six-figure deal showed zero LinkedIn Ads attribution in HubSpot. When I dug, I found 563 impressions, 20 engagements and 18 clicks on my TLAs before the deal even opened, and every big LinkedIn-influenced deal I checked had TLA engagement first.
The gap is 3 to 5x across the 150+ accounts Justin Rowe has checked, which is what you'd expect with an 11-person buying committee and 60 to 180 day cycles. Credit every campaign that touched the account before the deal was created. When I switched, my reported pipeline influence went up 3x+ with zero changes to the campaigns. Here is a ZenABM deals dashboard with campaign attribution:
Campaign Manager gives you impressions, clicks, CTR and CPC, but it can't tell you which companies are engaging or whether they're moving toward a purchase. Max's free tip is to segment by list in Campaign Manager to get per-company engagement, paid and organic. It's capped at a top-25 view and disconnected from your CRM, but it's something.
The full version needs the LinkedIn Ads API, which is what ZenABM reads: every company that saw, clicked or engaged, per campaign and per ad, mapped to your target list and joined to CRM deal events on one timeline. That makes revenue attribution per ABM campaign possible, with LinkedIn Ads, Google Ads, Reddit Ads, AI referrals and organic visits on one company journey.
This is a real reporting quirk almost nobody knows. Campaign Manager throws away all spend that didn't produce a conversion, then divides by what's left. AJ Wilcox, founder at B2Linked, found $3,793 spent and 2 conversions, a reported cost per conversion of $104 and a real number of about $1,896. LinkedIn support confirmed it's expected behaviour.
"By having the wrong calculation (that makes the performance look better than it actually is), LI is actually just confusing their advertisers" AJ Wilcox, on LinkedIn
So export the data and divide yourself.
Half the damage happens before the ad and after the click, so audit in the order things break. Patrick Cumming's order, which I follow, is tracking, then audiences from revenue data, then budget for reach and frequency, and creative last. Most audits start with creative, which is backwards.
If you want one command instead, ZenABM's /linkedin-abm-audit skill grades your live account across six dimensions and returns a fix list ordered by dollars at stake.
You don't need a strategy meeting, a new tool or a bigger budget for any of this. It's a skill issue, not a platform problem 🤷♀️. The other 52 will still be here next week.
I built these because I made the mistakes above. Here is which ZenABM capability does which step:
/linkedin-abm-audit, /budget-wasters and /ad-decay run complete analyses, and every change is confirmed before it reaches LinkedIn. Here is Zena answering a question about LinkedIn Ads mistakes:
The most common LinkedIn ads mistakes are targeting mistakes, because they're invisible: CTR and CPC look normal while you pay to reach the wrong people. Three show up in almost every audit. People skip the Audience Hub preview, they don't know LinkedIn expands job titles into broader "super titles", and they leave location on "Recent or Permanent", which includes tourists.
Use manual bidding in almost every case. Start 30% below LinkedIn's recommended range and check that daily spend reaches 95% of budget. Maximum Delivery charges per impression and spends as fast as possible, so the only exception, per Max Herczeg, is Thought Leader Ads on a Brand Awareness objective. Just don't bid so low that you only reach the cheapest slice of your audience.
Cold audiences need 6 to 10 impressions per person per month in total, and warm audiences need 10+. Per creative, replace it after 2.5 to 3 impressions per person, and TLAs start fading from about 2. Monthly audience penetration should be 30% to 50% for cold audiences and 70% to 90% for warm ones.
CTR doesn't predict pipeline. Across 33 companies with complete pipeline data, CTR correlates negatively with pipeline (Spearman rho -0.170) and spend correlates positively (0.566). A good CTR with no pipeline usually means Audience Network is inflating clicks, your targeting reaches the wrong people cheaply, or you're not spending enough to reach real buyers often enough.
Use Website Visits for traffic and demo signups, Engagement for in-feed content and retargeting pools, and Lead Generation for gated content. Avoid Website Conversions, which is a more expensive version of Website Visits. In Tim Davidson's test it cost $88 per conversion, while Website Visits cost $21.51 per landing page click (different metrics). Use Brand Awareness only for Thought Leader Ads when you want reach.
Mostly because of setup, not the channel. Common causes are audiences that are too big for the budget, default settings left on, a low match rate, last-click attribution that hides awareness ads, and no follow-up on engaged accounts. Fix tracking first, then audiences, then budget, and only then creative.