The account was spending $22K on LinkedIn for eight SQLs.
A B2B SaaS company, mid-market workflow product, $7M ARR, came in running $22,000 per month on LinkedIn Sponsored Content. The campaigns were technically well-built: tight job-title and seniority targeting, two account-based audiences for named-account lists, Conversation Ads for higher-intent prompts, and lead-gen forms set to capture work emails. The agency before us had built the program over 18 months and called it the company’s most mature paid channel.
Over the trailing 90 days, the program had produced 24 sales-qualified leads. Eight per month. Cost per SQL: $2,750. Average deal size: $24,000 ARR. Win rate on those SQLs: 18%. The math meant LinkedIn was producing roughly $43,000 in new ARR per month for $22,000 in spend, plus the agency retainer of $3,500.
The CFO had stopped asking why CAC was high. The CEO was asking why pipeline was thin. The marketing leader was defending LinkedIn on the strength of an attribution model that credited LinkedIn for any opportunity where a lead had ever filled a LinkedIn form, including leads that came in through inbound, partner, and content channels first.
This is the pattern. We see a version of it on roughly one in three B2B SaaS audits.
The audit found the obvious
The first finding was that LinkedIn’s view-through and engagement credit was being read as primary attribution by the marketing team. When we separated LinkedIn-originated SQLs from LinkedIn-touched SQLs (using the company’s own CRM stage-entry data, not the platform reporting), the originated count over the prior six months was 18, not 51.
The second finding was that the company’s best-performing LinkedIn ad concept, a comparison-focused carousel, was running unchanged for 11 months. Engagement had been declining month over month for the last six. The creative library across all campaigns was four assets total.
The third finding was that the company’s Google Search account was running $4,200 per month, producing 19 SQLs per month at $740 each, with the budget capped because the agency thought Google Search was “saturated.” Looking at the actual impression share data, the campaigns were impression-share-lost-budget at 64%. The cap was self-imposed.
The fourth finding was that the company had no Meta presence at all. The marketing leader’s stated reason was “we are B2B, our buyer is not on Meta.” The buyer demographic (ops and engineering managers, age 30 to 50) overlaps Meta’s daily active user base at roughly 95%.
The reallocation
The recommendation was to kill LinkedIn entirely for one quarter and move the $22,000 across Google Search and Meta Reels.
This was not popular. The marketing leader had built a six-quarter narrative around LinkedIn as the company’s flagship channel. The agency relationship had pricing tied to LinkedIn complexity. The CEO had been told for two years that LinkedIn was the only credible B2B paid channel. The CFO was the only one in favor of the test, on the simple math that $43K of ARR for $25K of spend including retainer was, after sales and service costs, near zero contribution margin.
The compromise: kill LinkedIn for 90 days, run the test, hold the team’s compensation reviews on hold for the quarter so no one’s bonus depended on the channel mix.
The reallocation:
$16,000 per month to Google Search, lifting the existing campaign’s budget from $4,200 to $16,000 with no other changes. Impression share lost (budget) projected to drop from 64% to roughly 18% based on the auction insights data.
$6,000 per month to a new Meta Reels program. Three creative concepts produced in-house by the company’s product marketing team, six variants of each, optimized against a Lead event tied to a 15-minute product demo signup. No interest targeting; Advantage+ defaults.
LinkedIn paused, with the LinkedIn page itself kept active for organic and sales-team posting.
What happened in 90 days
Total monthly spend held at $22,000. The LinkedIn agency retainer of $3,500 came off the books since the work shifted to channels we managed. Net monthly outlay including our retainer: roughly $24,000, slightly above the prior $25,500.
Google Search produced 47 SQLs per month at $340 each. The impression-share-lost (budget) dropped to 14% with the new spend level. Win rate on those SQLs was 24%, six points higher than the LinkedIn SQLs, because the searches the campaign was capturing were higher-intent than the upper-funnel LinkedIn interactions.
Meta Reels produced 21 SQLs per month at $285 each after a six-week ramp. Win rate was 19%, roughly in line with LinkedIn. The Reels concept that worked best was a 22-second product walkthrough scripted from a customer success call recording.
Total SQLs per month rose from eight to 68. Total pipeline created rose 31% in two quarters versus the prior two quarters. CAC dropped from $13,400 to $4,800.
Why the previous setup failed
Three failures, all common.
The first was channel attribution credit confusion. The marketing team was reading LinkedIn as their best channel because LinkedIn’s reporting credited it for any deal where a lead had ever interacted with a LinkedIn ad. When we built attribution from CRM stage-entry data (the first source that produced a lead with their work email in the company’s system), LinkedIn’s originated share fell from 51% of SQLs to under 20%. The channel was getting credit for influence on leads that had originated elsewhere.
The second was creative fatigue going unaddressed. The same four LinkedIn ads had been running for nearly a year. LinkedIn’s audience is smaller and more concentrated than Meta’s; the same buyer sees the same ad five or six times in a quarter on LinkedIn before they see it twice on Meta 1. Stale creative on LinkedIn is more costly than stale creative on Meta because the rotation depth is shallower.
The third was channel allocation built on a stereotype, not on math. The “our buyer is not on Meta” framing is one of the most expensive sentences in B2B SaaS. Meta’s audience overlap with most B2B buyer personas is high; CPM on B2B Meta Reels was $11 in Q1 2026, down 22.6% year over year 2. The platform’s targeting precision for B2B has improved meaningfully through algorithmic optimization; Reels-format creative reaches working professionals consistently.
The pattern this illustrates
The firm’s position on diversification, written up at length in the more-channels piece, is that concentration produces better data, faster learning, and more reliable optimization. This account is one instance.
The deeper pattern is that B2B SaaS marketing leaders often defend LinkedIn out of a category narrative rather than from the math. LinkedIn ads can work; the firm has built profitable LinkedIn programs at clients with the right combination of ICP precision, deal size, and creative production volume. They cannot work at $2,750 per SQL with a four-asset creative library at $7M ARR.
Median LinkedIn CPC for B2B SaaS sits between $12 and $18 in 2026 3. For an account with ACV under $40K and a sales cycle under three months, the LinkedIn-CPC-to-CAC math rarely closes. For accounts with ACV over $100K and named-account ABM motions, it sometimes does.
Most B2B SaaS accounts under $20M ARR fit the first profile, not the second. Their LinkedIn spend earns out only when the attribution model is generous. Tighten the attribution and the channel often does not survive its own math.
What this is not saying
This is not a claim that LinkedIn is a bad channel. It is a claim that LinkedIn is a channel that requires specific conditions to earn its cost, and that those conditions are uncommon at the budget and ACV profile most mid-market SaaS companies operate at.
It is also not a claim that Meta works for every B2B SaaS account. Meta Reels worked here because the company had product marketers who could produce ten short videos per month and a product walkthrough that translated to a 22-second hook. A B2B SaaS account selling a $250K-ACV product to enterprise CIOs would likely not see the same lift; the buyer set is too narrow for Meta’s targeting to find efficiently.
The lesson is the math. The account had been operating on a channel-mix story that was not supported by the originated-attribution data. When the story and the math disagreed, the math was right. The right move was to follow the math, not the story.
The work after the test
The 90-day test ended with the marketing team running Google Search and Meta Reels as primary channels, and with LinkedIn returning later in the year for an ABM-only motion targeting 80 named accounts at the enterprise tier. The retainer-driven approach was replaced with a budget-driven one, and the agency model with it.
The CFO stopped asking why CAC was high. The CEO stopped asking why pipeline was thin. The marketing leader updated the channel narrative and ran the next planning cycle without LinkedIn as the flagship.
The right answer was usually “spend less, on fewer things.” In this case it was “spend the same, on different things.” Same principle. Different application. The leverage sits in matching the channel to the math, not in defending a channel because it is the one the category talks about.