Make confident, data-driven decisions with actionable ad spend insights.
13 min read
LinkedIn is unique in the paid media landscape. Unlike platforms geared toward immediate e-commerce transactions (B2C), LinkedIn is purely a B2B ecosystem focused on high-value, high-friction conversions: qualified leads, MQLs, SQLs, and ultimately, signed enterprise contracts. Consequently, the way you calculate, benchmark, and optimize Return on Ad Spend (ROAS) must fundamentally shift.
Orla Gallagher
PPC & Paid Social Expert
Last Updated
November 26, 2025
You’ve been asked the question in a marketing meeting. You’ve probably asked it yourself. "What's a good ROAS for our LinkedIn campaigns?"
You get a vague answer. "It depends." Or maybe someone pulls up an industry blog post with a colorful chart showing that a 4:1 ROAS is "average" for B2B tech. Everyone nods, feeling a little more certain.
That certainty is a fantasy.
The entire conversation is built on a foundation of sand. The ROAS number you see in your LinkedIn Ads dashboard is, to put it bluntly, a guess. It’s a metric so distorted by technical limitations and platform self-interest that using it to make critical budget decisions is like navigating a minefield with a pogo stick.
The real question isn't "What's a good ROAS?" It's "Can I even trust the ROAS number I'm looking at?" For most B2B advertisers, the answer is a resounding no.
Before you fire your agency or blame your marketing team, you need to understand the systemic issues that are warping your data before it ever reaches your dashboard. This isn't about bad ad copy; it's about a broken data pipeline.
Let's be clear: LinkedIn wants to show you a high ROAS. Its business model depends on you believing your ad spend is effective. To do this, they use generous attribution models.
A "view-through" conversion, for instance, gives LinkedIn credit if someone saw your ad, didn't click, but converted on your site days later. Did the ad cause the conversion? Or was it the webinar they attended last week and the email they got this morning? The platform will always vote for the ad.
This creates a "walled garden" effect. LinkedIn is incentivized to take credit for any positive outcome, making it nearly impossible to discern its true incremental impact. You're not getting an objective report; you're reading the platform's own performance review.
Here is the technical reality that most "how-to" guides ignore. The LinkedIn Insight Tag, the piece of code that tracks conversions, is a third-party script.
Why does this matter?
Because a significant portion of your target audience, especially in tech-savvy B2B sectors, uses ad blockers. Estimates put adoption at 30-40% and rising. These tools block the LinkedIn Insight Tag by default.
Then there's Apple's Intelligent Tracking Prevention (ITP). It’s built into Safari and aggressively limits the lifespan of third-party cookies, making it difficult to track user journeys over time.
The result? A huge chunk of your most valuable prospects see an ad, click to your site, and convert. But because the tag was blocked, the conversion is never reported back to LinkedIn. Your 'Return' is artificially lowered, and your ROAS plummets.
The internet is flooded with non-human traffic. LinkedIn is no exception. Bots and click farms constantly crawl the web, clicking on ads and sometimes even filling out forms with fake information.
This fake engagement directly inflates your costs. Every dollar spent on a bot click is a dollar added to the 'Ad Spend' part of your ROAS calculation with zero chance of generating revenue.
While your 'Return' is being under-reported due to blockers, your 'Spend' is being over-reported due to fraud. This combination is toxic for your ROAS calculation.
Your ideal customer profile might be a CTO at a Fortune 500 company in the United States. But that CTO is likely using a corporate VPN that routes their traffic through a server in another country.
Your campaign data might show a surge of traffic from a region you weren't even targeting, leading you to believe your targeting is broken. In reality, your targeting might be perfect, but the data is being masked. This makes it impossible to trust the geographic and firmographic reporting inside LinkedIn.
This isn't just a reporting headache. Flawed data leads to terrible business decisions that echo across departments.
Imagine your team launches a campaign targeting senior cybersecurity professionals. The LinkedIn dashboard shows plenty of clicks but almost zero attributed conversions. Following the data, you conclude the campaign is a failure and shut it down.
The hidden reality: The campaign was a massive success. Dozens of security-conscious targets clicked and requested demos, but their browsers all blocked your tracking pixel. You just killed your best-performing campaign of the quarter because you were relying on incomplete data.
Now, consider the opposite scenario. You run a LinkedIn Lead Gen Form campaign. The cost-per-lead (CPL) is fantastic, and you send a hundred "MQLs" over to the sales team.
The sales reps start making calls, only to find that the leads are students, competitors, or simply don't exist. The forms were filled out by bots or low-intent users who clicked a button without a second thought. Sales wastes weeks chasing ghosts, and their trust in marketing's ability to generate qualified leads is shattered.
The CFO looks at your budget request. You've spent $50,000 on LinkedIn this quarter. The platform-reported ROAS is a dismal 1.5:1. From their perspective, you're barely breaking even.
You feel the campaigns are driving real pipeline, but you can't prove it. The attribution data is a mess of holes and contradictions. You can't defend the spend, the budget gets cut, and a critical growth engine is throttled.
The internet is full of well-meaning advice for improving LinkedIn ROAS. The problem is that most of it focuses on optimizing a system that is fundamentally broken.
This is standard advice. And yes, good creative and smart bidding are important. But when you're optimizing based on data that is missing 30% of your conversions and is polluted with bot clicks, what are you actually optimizing for?
You might be "optimizing" away from the ad that appeals most to senior decision-makers using Safari and towards the ad that performs best with bots and junior employees. It's like trying to win a race by polishing the hubcaps on a car with a faulty engine.
As we've seen, this is a dangerous trade-off. You solve the off-platform tracking problem by keeping the user inside LinkedIn's garden. But in doing so, you lower the barrier to entry so much that you invite a flood of low-quality and fraudulent submissions. You trade a tracking problem for a massive lead quality and sales alignment problem.
UTMs are essential hygiene for tracking clicks, but they are a crude tool for measuring true ROI. They only tell you about the last click, completely ignoring the complex, multi-touch journey of a typical B2B buyer. They do nothing to solve the problem of view-through attribution or, more importantly, the conversions that were never recorded in the first place due to blocked trackers.
If you can't trust the platform's data, the only solution is to build your own source of truth. This means shifting your entire mindset from relying on third-party pixels to owning your data from the first touch to the final conversion.
As Christopher S. Penn, Chief Data Scientist at TrustInsights.ai, aptly puts it, "Platform-reported metrics are marketing. Your server-side data is accounting. Trust the accounting."
The core of the solution is to stop using trackers served from third-party domains (like linkedin.com) and start serving them from your own domain.
This is done by setting up a simple CNAME DNS record to route a subdomain (e.g., analytics.yourdomain.com) to a dedicated tracking server. When your website's tracking script loads from your own subdomain, browsers and ad blockers see it as a "first-party" resource. It's treated as a trusted and essential part of your website, not as an intrusive foreign spy.
This single change allows you to bypass the vast majority of ad blockers and ITP restrictions, instantly recovering the 30-40% of conversion data you were losing.
Capturing all the data is the first step. The second is cleaning it.
A modern first-party data infrastructure doesn't just collect data; it validates it. Before a "click" or "conversion" event is even recorded, the system should automatically filter out traffic from known bots, data center IP addresses, and malicious user agents.
You're left with a clean, verified log of real human interactions. This is your ground truth.
Once you have this clean, complete dataset, you can use it to educate the ad platforms. Instead of relying on the faulty browser-based pixel, you send your verified conversion data directly from your server to LinkedIn's server via their Conversions API (CAPI).
This achieves two critical goals:
This table illustrates the profound difference in approach:
| Metric/Process | The Old Way (Relying on LinkedIn Pixel) | The New Way (First-Party Data Pipeline) |
|---|---|---|
| Data Capture | LinkedIn Insight Tag (3rd-party) blocked by ~30% of users. | First-party script captures 99%+ of sessions. |
| Traffic Quality | Clicks include bots, competitors, and junk traffic. | Bots and fraudulent traffic are filtered out pre-analysis. |
| Conversion Data | Missing conversions from users with blockers/ITP. | Complete conversion data is captured on your server. |
| ROAS Calculation | Inflated 'Ad Spend' (from junk clicks) / Under-reported 'Return' (from missed conversions) = Artificially Low ROAS | Clean 'Ad Spend' / Accurate 'Return' = True ROAS |
| Optimization Signal | You optimize based on a skewed, incomplete picture. | You optimize based on verified human behavior. |
We're back to the original question, but now we can answer it intelligently.
With a clean data pipeline, you can finally stop chasing vague industry benchmarks. As one VP of Growth put it, "Industry benchmarks are a comforting lie. The only benchmark that matters is the one that proves to your CFO that marketing is a revenue driver, not a cost center."
A "good" ROAS is one that is positive, improving, and accurately reflects the economics of your specific business. If your average contract value (ACV) is $150,000 and your sales cycle is nine months, a 3:1 ROAS on LinkedIn might be phenomenal. If you're selling a $1,000 per year subscription, you might need a 7:1 ROAS to be profitable.
The point is, once you trust your numbers, you can have this strategic conversation with confidence. You can build a real financial model for your marketing, not just a slide deck with charts from LinkedIn's dashboard.
Moving from guesswork to certainty requires a deliberate change in strategy and tooling. Here’s how to start.
Look at your current web analytics. What percentage of your traffic comes from the Safari browser? That number is a good proxy for the minimum amount of data you're losing to ITP. Now add another 20-30% for other browsers with ad blockers. The result is your data blind spot. Is it a number you're comfortable with?
Where do you go to see if your marketing is working? If the first answer is the LinkedIn Ads dashboard, you have a problem. Your primary source of truth should be your own analytics platform, one that you control and that reflects your complete, unfiltered reality.
Stop trying to patch a leaky bucket. It's time to replace the bucket. The technical challenges of building and maintaining a first-party data pipeline from scratch are significant.
This is why integrated solutions like DataCops exist. They are designed to solve this exact problem. By consolidating tracking into a single, first-party script that you control, you instantly bypass ad blockers and ITP. The system automatically filters out fraud and bot traffic. It then sends this clean, complete conversion data back to platforms like LinkedIn via their CAPI.
It’s not about gaming the system. It's about getting an honest accounting of your marketing efforts so you can finally make decisions with confidence.
This isn't just about LinkedIn. The entire digital advertising ecosystem is being rebuilt around privacy and first-party data. With Google's phasing out of the third-party cookie, businesses that haven't established their own data foundation will be flying completely blind.
Mastering your data pipeline is no longer a competitive advantage for optimizing ROAS; it is becoming the fundamental requirement for participating in digital marketing at all. The time to get your data house in order was yesterday. The next best time is now.
No. This is about the critical distinction between technical blocking and user consent. First-party tracking still operates entirely within the bounds of user consent. A proper solution includes a TCF-certified First Party Consent Management Platform (CMP) to ensure you honor every user's choice regarding data processing. The goal is to get accurate measurement from users who have consented, but whose browsers are technically blocking your trackers. It restores data from technical gaps, it doesn't override explicit user denial.
You can attempt a piece of it, but it's a far more complex and incomplete solution. GTM server-side requires significant, ongoing engineering resources to set up and maintain. Crucially, it does not have built-in fraud and bot detection. You would be responsible for building that logic yourself, or you would simply be passing the same dirty data through a different pipe. An integrated platform like DataCops is designed to handle the first-party data capture, fraud filtering, consent management, and CAPI integrations in one unified, managed solution, providing a much cleaner signal with far less effort.
A high ROAS reported by the platform can be just as misleading as a low one. Is it real, or is it inflated by generous view-through attribution for conversions that would have happened anyway? Are you optimizing toward a segment that looks good on paper but has a low close rate for sales? Without a verified, first-party source of truth to compare against, you can't be sure your "great" ROAS is based on actual, incremental revenue. A clean data pipeline validates your wins, protects you from false positives, and ensures your budget is truly driving the bottom line.
