Josh Silverbauer: Why Server-Side Tagging Stopped Being a Nice-to-Have

Roughly 30 percent of conversions vanish from the reports of companies that still rely on client-side tracking alone, and most marketing leaders have no idea it is happening. They are reading dashboards, calculating return on ad spend, and allocating budget against a number that is structurally wrong. Josh Silverbauer, who audits digital analytics architectures for a living, traces the problem back to a decision made outside any marketing department: browsers, led by Apple with Safari on macOS and iOS, began curbing the ability to send data directly back to third-party platforms in the name of privacy. That choice turned a technical preference into a commercial liability. Server-side tagging was once the kind of project an analytics team pitched and lost. It is now the difference between measuring a marketing program and guessing at it.

The Data You Are Already Losing

The mechanics are simple enough that executives should not be able to hide behind their complexity. When a user moves from one site to another and the sending site fires a request straight at a platform like Facebook, iOS blocks it. The conversion still happened. The record of it never arrives. “You’re missing probably about 30% of your conversions if you rely on client-side tracking alone,” Silverbauer says. That gap does not distribute itself evenly or politely. It lands on whichever channels serve the most privacy-restricted traffic, which means the campaigns getting cut are not necessarily the ones performing worst.

Server-side tagging changes the destination rather than the data. Instead of firing at facebook.com, a company sends the event to a subdomain it owns, sitting on its own server, then forwards from there. “It looks like I’m sending data to myself,” Silverbauer explains. Ad blockers, built to intercept third-party requests, have nothing to catch. He stresses that consent remains essential. “This is not a way to get around consent,” he says. “If people don’t want to be tracked, they can say I don’t want to be tracked.” His objection is to a different authority entirely. “Nobody’s marketing should rely on an external browser saying you can or can’t do this. Like that’s silly.” Safari behaves one way, Chrome another, and no business should let that inconsistency define its measurement.

Why Nothing Ever Lines Up, And Why That Is Fine

The second argument runs against a belief most finance and marketing leaders hold without examining it: that platform numbers and customer relationship management (CRM) system numbers should eventually agree. They will not. “All the platforms are greedy,” Silverbauer says. “They want you to spend more money. So they are always going to report higher than what your CRM says.” The CRM is usually the source of truth. The ad platforms are optimization layers, useful for steering spend inside their own ecosystems and useless as arbiters of total performance. Treat each one accordingly, pick a single attribution system as the baseline, and judge the rest on whether they move directionally with it. “If it does match, then it’s a miracle.”

What matters is knowing when a discrepancy signals a break rather than the normal gap. Finding the break is the hard part. “One of the trickiest parts of tracking and tagging infrastructure is finding where in the sequence of configured components something has broken,” Silverbauer says. Someone swaps out a form, and a tag referencing an ID on that form stops firing. A template updates. The ads platform itself updates. “There’s a lot of places to go treasure hunting for why tracking might be broken.” Organizations without a documented chain, from website to tag manager to platform, will spend weeks on that hunt every time, and will have been optimizing against bad numbers for the duration.

The Case To Finance, And The Skill That Survives AI

Marketing directors asked to fund infrastructure they will never see or touch need the argument translated into money, and the translation is not difficult. More complete data produces better platform optimization, better targeting, and better audience building. Richer signals sent server-side, including identifiers like email and phone number, let a platform match users without any pixel dependency at all. The strongest version of this is feeding events directly from the CRM, which Silverbauer calls the most reliable approach, partly because it carries real identity data and partly because it reaches further down the funnel. With enough volume, a company can optimize toward qualified leads or closed sales rather than raw leads. That is a different quality of marketing program, not a cleaner version of the same one.

Which raises the question of what analytics people should be good at now. Silverbauer thinks the technical implementation layer is already on borrowed time. “AI is capable enough to set up tracking,” he says. “It’s able to look through an entire site, and you’re going to be able to be like, I want to track this, this, and this, and it can do that.” What it cannot supply is judgment about what is worth tracking in the first place. “Most businesses lack the data maturity to understand what’s valuable to them. They might have a few ideas, but they can’t articulate it.” The practitioner who survives is the one who sits with an organization, puts business objectives first, defines the key performance indicators underneath them, and designs the architecture that produces the answer. “It’s more conversational, it’s more thinking, it’s more architecture strategy as opposed to technical implementation,” Silverbauer says. “You now have these arms that anybody can use to do technical implementation. What really will make you stand out is your brain.”

Follow Josh Silverbauer on LinkedIn for more insights on server-side tagging, measurement strategy, and analytics architecture.