ABM Mistakes: Why $75K B2B Spend Failed in 2026

Account-Based Marketing is supposed to be about precision, but campaigns tank all the time when the social targeting is wrong. You end up just burning cash and missing B2B lead generation opportunities. We just tore down a B2B SaaS campaign that shows all the classic ABM mistakes in social targeting, specifically on LinkedIn and Meta. How can a huge budget produce such weak results? Simple: the audience wasn’t really there.

The Campaign: A Deep Dive into Disconnects

The campaign we analyzed, “Project Ascent,” was supposed to pull in qualified leads for an AI-powered analytics platform built for the financial services industry. The client, a mid-sized B2B SaaS company, threw a pretty big budget at it: $75,000 over 10 weeks in Q3 2026. The goal was pretty aggressive, too. They wanted 50 sales-qualified leads (SQLs) and were aiming for a Cost Per Lead (CPL) of $500 or less, plus a 1.5x Return on Ad Spend (ROAS) based on their average LTV.

Strategy and Initial Approach

The strategy was to go after decision-makers and influencers inside specific financial institutions, banks, hedge funds, asset management firms, in the US and UK. The team went all-in on LinkedIn for its pro targeting and used Meta (Facebook and Instagram) mostly for retargeting and wider reach. Their whole plan was built around thought leadership content and case studies.

• Target Audience: VPs of Risk Management, Heads of Data Analytics, Chief Investment Officers.
• Platforms: LinkedIn Ads, Meta Ads (Facebook & Instagram).
• Content Pillars: Whitepapers on AI in risk assessment, webinars on predictive analytics, success stories from early adopters.
• Primary CTA: Download a whitepaper / Register for a webinar.

Creative Execution: One Size Fits… Few

The creative had that glossy, corporate feel. All the ads used slick stock photos of diverse professionals looking smart in modern offices, with headlines like “Unlock Deeper Insights” and “Transform Your Data.” They looked good, sure, but they weren’t specific at all. The same basic ads were pushed out to every target segment, with only tiny copy tweaks for each platform.

On LinkedIn, they ran carousel ads to show off platform features and single image ads to push whitepaper downloads. Over on Meta, they used short (under 30 seconds) video ads showing the UI, then hit people again with retargeting image ads asking them to book a demo. The client thought the creative was a great reflection of their brand, but as we saw, looking good doesn’t mean you’ll get conversions.

The Data Unveiled: What Actually Happened

Five weeks in, the campaign numbers were bad. Here’s a quick look at the metrics:

The campaign was getting tons of impressions, which means the ads were definitely being served, but the garbage CTR told us nobody cared. The CPL for an actual SQL was more than four times their target, a huge red flag for inefficiency. This is exactly what happens when your targeting and messaging are completely out of sync, classic ABM mistakes.

Social Targeting Pitfalls: Where It Went Wrong

Our post-mortem pointed to a few critical errors in the social targeting, especially on LinkedIn. The team had leaned on broad job titles and industry categories, thinking that would be enough to hit the right people.

1. Overly Broad LinkedIn Targeting: The first LinkedIn audience included titles like “Financial Analyst” and “Portfolio Manager” but didn’t filter for seniority. These people work with data, but they almost never have budget authority like a VP or CIO. The campaign also targeted a list of 500 companies but didn’t bother to segment by department, so ads were getting served to totally irrelevant employees. According to a recent IAB B2B Digital Ad Spend Report, you can get a 35% lift in engagement just by using precise account-level targeting instead of going broad.
2. Lack of Intent Signals: They weren’t using any behavioral or intent data. While LinkedIn lets you target based on groups, skills, and seniority, this campaign just used static firmographics. They didn’t integrate with any intent data platforms (like Bombora or G2 Buyer Intent) to see which accounts were actually out there looking for AI analytics tools. This meant ads were shown to a lot of people who weren’t in a buying cycle which blew up the impression count without generating any real leads.
3. Meta Audience Mismatch: On Meta, the plan was to create lookalike audiences from website visitors and CRM lists. The problem was, the CRM list they used was full of contacts who weren’t decision-makers or worked at smaller companies outside their target enterprise segment. That totally diluted the lookalike audience, so ads ended up in front of people who were never going to convert. On top of that, the creative was the same boring corporate stuff they used on LinkedIn, completely ignoring the fact that people use Facebook and Instagram differently (and more casually).
4. Insufficient Exclusion Targeting: The campaign was missing strong exclusion lists. They kept serving ads to their own competitors, current customers, and people who had already downloaded the whitepaper. It was a straight-up waste of money and probably annoyed a few people.

Optimization and Course Correction

The performance was so bad we had to jump in and start optimizing fast for the remaining five weeks. We had to fix a few things at once: the targeting, the creative, and the conversion funnel.

Targeting Refinement: Precision Over Volume

The biggest fix was a complete overhaul of the targeting. On LinkedIn, we:

• Implemented Seniority Filters: We limited job titles to “Director” and above, focusing specifically on roles like “Head of Risk,” “Chief Data Officer,” and “VP of Quantitative Analysis.”
• Layered Skills and Groups: We added layers for specific skills (like “Machine Learning in Finance,” “Algorithmic Trading,” “Regulatory Compliance”) and membership in relevant professional groups.
• Introduced Account-Based Retargeting: We built custom audiences from a hand-picked list of 100 target accounts, zeroing in on employees who had visited specific product pages on the client’s site.
• Integrated Intent Data (Pilot): We carved out a small part of the budget to test a third-party intent data provider, targeting companies that were actively searching for “AI risk analytics” terms.

For Meta, we:

• Refined Lookalike Seeds: We rebuilt the lookalike audiences from scratch, this time using only confirmed SQLs from old campaigns and people who had seriously engaged with high-value content (like watching 75% of a webinar).
• Leveraged Custom Audiences for Retargeting: We created super-granular custom audiences of website visitors based on which pages they’d seen (like the pricing page or demo request page) so we could hit them with perfectly relevant retargeting ads.

Creative Overhaul: Speak to the Pain

We threw out the generic creative and replaced it with content that actually spoke to their audience’s problems. We developed:

• Persona-Specific Ad Copy: For VPs of Risk, our headlines were about “Mitigating Unseen Financial Risks.” For the Heads of Data Analytics, we talked about “Simplifying Data Pipelines for Predictive Accuracy.”
• Visual Variety: We got rid of the boring stock photos and started using data visualizations, actual screenshots of the dashboard, and short testimonial clips (with permission, of course).
• A/B Testing CTAs: We tested “Download the 2026 Risk Report” against “See a Personalized Demo” to see what different segments actually wanted to do.

Landing Page Optimization

The funnel itself was leaky. We tightened up the landing pages by cutting the form fields from 8 down to 5 and making sure the value proposition on the page matched the ad they just clicked. We also embedded a short explainer video on the main landing pages to show people what the platform did, fast.

The Turnaround: Improved Performance

These changes made a huge difference in the campaign’s final five weeks. They still didn’t hit that ambitious SQL target, but the campaign’s efficiency shot way up.

The CTR more than doubled, which tells you the new creative and targeting were hitting the mark. Sure, total impressions dropped, but the quality of those impressions was way better, which is why the CPL plummeted and the ROAS started to recover. They didn’t hit the original $500 CPL target for SQLs, but cutting it in half in just five weeks shows how powerful quick optimization can be. With ABM, precision beats volume every time, no matter how big your budget is.

Lessons Learned: Avoiding Future Pitfalls

What this whole teardown shows is that ABM mistakes in social targeting happen when you get lazy and rely on broad assumptions. My advice is to always start hyper-specific with your targeting, even if that means your initial audience feels small. You can always broaden it later if things are working. A huge audience doesn’t guarantee results. It usually just guarantees you’ll waste money. Is there a better way? The real win comes from hitting the right people with a message that solves their problem right when they need it. That means you have to constantly refine your audience segments, test your creative like crazy, and pull in every bit of data you can, from your CRM to intent signals, to build a truly smart targeting strategy. A recent eMarketer report on B2B Social Media Trends really drives home how important AI-driven audience segmentation is becoming for getting good ROI.

For more on optimizing your B2B social game, check out our guide on B2B Social Algorithms: 2026 Strategy Shift for 40% ROI. Getting a handle on these shifts can make or break your campaign’s performance. You might also want to see how AI is changing B2B lead qualification on social media, which is opening up new ways to be more precise and efficient. And to make sure your ABM is as personal as possible, look into the new breakthroughs in AI ABM personalization that are defining the next wave of targeted marketing.