LinkedIn now accounts for 64% of B2B social media conversions, up from 51% in 2024. And 77% of B2B marketers plan to increase paid social spend over the next twelve months. Those numbers tell you where the money is going. They don't tell you what's working once it gets there.

Eight demand gen operators shared what they ran, what they spent, and what came back as pipeline. Documented plays with named people and real numbers. Here's what each one teaches about where B2B creative and channel strategy sits right now.

Meme Creative on LinkedIn Broke Through CPM Fatigue

Alexander Goodwin, Director of Demand Generation at Fingerprint, was bidding against every competitor running the same fear-based fraud-detection creative on LinkedIn. Same audience, same message, rising CPCs. He swapped product value props for copy that called out buyer behavior directly: "Scrolling won't stop fraud attacks. Fingerprint will." Thirty creative variants over two months. Signups jumped 60% in month one.

The lesson isn't "use memes." Creative fatigue is a signal problem, not a bidding problem. When everyone in a category converges on the same tone, the cheapest way to win attention is to break the pattern.

TV on a $20K Budget Produced $200K in Pipeline

Hooman Javidan-Nejad, Senior Director of Marketing at Tatari, repurposed existing video for TV, including a 20-minute sponsored fireside chat that cost under $1,000 in post-production. Total spend: $20,000. Result: 2 million impressions and nearly $200K in pipeline. The trade-off: measurement is harder than digital, and you're relying on directional attribution unless you instrument geo-holdouts or matched-market tests.

OOH with Incrementality Testing

SurveyMonkey's problem wasn't awareness. The problem was perception: buyers assumed all they did was surveys. Stefanie Wlotzki, Senior Paid Media Strategist, ran their "What if you just knew?" OOH campaign in Chicago and Boston as test markets against matched controls. Results: 244 million impressions, 7% lift during a period (Easter) that normally dips.

Her warning: OOH impact can take six to nine months to show up in the metrics that matter. If you need pipeline this quarter, this isn't your play. If you need to shift brand perception over two or three quarters, the holdout design is what makes it defensible to the CFO.

A 9-Minute Video Ad That Cut Cost Per Demo 60%

Jake Newby, Director of Growth Marketing at Lendio, ran a 9:49 explainer video targeted at named accounts on LinkedIn. Roughly $4,500 in spend produced 5 booked demos, with cost per demo 60% below Lendio's LinkedIn average. Platform engagement metrics would punish a video this long. But Jake measured cost per booked demo, not completion rate. Only the genuinely interested watch nine minutes. That's a feature, not a bug.

Incentivized Demos: Brand Teams Hate It, Pipeline Loves It

Anthony Blatner, Managing Director at Speedwork, offered gift cards in exchange for product demos. One enterprise SaaS company spent $100,000 against a specific account list and generated almost $4 million in pipeline. A separate HR company ran the same tactic as a competitor conquest play and saw 45x return on pipeline to spend. The guardrail is targeting: run it against a tight account list, not a broad audience.

ChatGPT Ads: Early Arbitrage, Real Caveats

Dylan Wingrove, Associate Director of Demand at Pacvue, pulled budget from Google and moved it into the ChatGPT Ads beta. One ad per prompt means zero competitor adjacency. His team increased spend 900% in four weeks. Advertiser counts on ChatGPT rose from about 300 in April 2026 to more than 820 by July 2026, and Demandbase reported ChatGPT referrals to tracked B2B websites jumped 303% year over year to 2.6 million monthly visits in June 2026.

The channel is real. But one VP Marketing reported spending $50K on ChatGPT ads and netting nothing because their audience skewed toward free-tier users. The diagnostic before you move budget: does your ICP actually use ChatGPT in a commercial research context?

Google AI Max with Guardrails

Kenna Rooney at EasyLlama tested Google's AI Max on an already-strong campaign, kept it within the existing daily budget, and restricted expansion to high-intent search terms. Within three weeks: qualified leads up 166%, CPA down 15%, pipeline generation up 3X. She didn't hand Google a blank check or let it expand into broad match. She constrained the experiment so the downside was bounded.

Soap-Opera Video for HR Software

Shlomo Genchin, Creative Director at Unbore.com, borrowed the Apple-vs.-Microsoft villain/hero structure for Hi Bob's HR platform. The format was a soap-opera breakup scene. Results: 1,000+ leads, 80% cheaper than their usual cost per lead, CPM down 50%. The principle underneath: narrative structure (villain, hero, relatable scenario) gives creative a skeleton that's easy to test and iterate on.

The Pattern Across All Eight

Every one of these plays worked for a specific team, in a specific context, against a specific problem. The common thread: each operator picked a conversion definition that mapped to pipeline, not vanity metrics, and constrained the experiment enough to read it cleanly. That's the part worth stealing, regardless of which play you run next.