Why we don't run ads for clients who don't have their tracking in order yet

A client came to us recently wanting to start paid ads the same week we kicked off their new site. We said no, not because we didn't want the work, but because their analytics setup couldn't yet tell us which visitors were converting into anything, and running ad spend against that gap felt like burning money to produce a number neither of us could trust.

This isn't a rare situation. It's close to the default state we find businesses in when they first come to us wanting to run paid traffic, and it's the single biggest reason ad budgets get wasted in a business's first few months of trying performance marketing at all.

It means a conversion event, a form submission, a purchase, a booking, fires reliably and gets attributed to the specific channel and even the specific ad that drove it. It means the numbers in the ads dashboard roughly match what the business actually experiences in leads or sales, within a reasonable margin, not off by a factor of three in either direction.

It sounds basic, and it is basic, which is exactly why it gets skipped. Setting it up correctly takes real time before a single dollar goes toward media spend, and that delay feels like friction to a client eager to see results. It is not friction. It's the only thing that makes the results that follow mean anything at all.

The most common failure mode is a client optimizing toward the wrong signal, because the platform's own reported conversions don't match reality, without either of us realizing it for weeks. Budget gets shifted toward campaigns that look like they're working in the dashboard and away from ones that are actually working in the business, and by the time the mismatch surfaces, real money has already gone the wrong direction.

The second failure mode is subtler: a campaign genuinely works, but nobody can prove it, so a client loses confidence and pulls the budget right when it was starting to pay off. Without solid tracking, a good result and a good story about a result look identical, and it's the story that tends to win when a client is deciding whether to keep spending.

A conversion event for every meaningful action on the site, tested manually by actually completing it ourselves and confirming it registers correctly, not just assumed to work because the code looks right. Server-side tracking as a backup where possible, since browser-based tracking alone increasingly misses a meaningful share of real conversions due to privacy settings and ad blockers.

We also set a baseline conversion rate from existing organic traffic before turning on paid traffic at all, wherever there's enough existing traffic to establish one. Without that baseline, there's no way to tell whether paid traffic is converting better, worse, or about the same as what the business was already getting for free.

A week or two spent fixing tracking before spend starts is a week or two the business isn't yet paying for clicks. Compared to weeks or months of spend optimized against broken numbers, the delay is cheap. We've had clients push back on this timeline more than once, and in every case where we held the line, they later agreed it was the right call.

We frame this explicitly in proposals now: tracking setup is a line item, not a footnote, with its own timeline and its own sign-off before any ad account gets funded. Clients who understand why upfront are far less likely to push for a shortcut later, once they've already seen the reasoning laid out plainly.

We don't lead with the technical explanation, because most clients don't care about the mechanics of server-side tracking. We lead with the plain version: if we can't tell you which ads actually worked, we're both flying blind with your money, and neither of us will know what to do more of next month.

Framed that way, almost every client agrees to the delay once they understand what it's protecting against. The ones who don't are usually the ones under the most external pressure to show activity fast, and those are exactly the clients most likely to get burned by spending against numbers that weren't real.

We open the client's analytics dashboard together, live, on the call, and walk through the last thirty days of reported conversions against what the client actually remembers happening in the business over that period. The gap, when there is one, is usually obvious within a few minutes, and it's a far more convincing argument than any slide we could prepare in advance.

This live audit has ended more than one "let's just start running ads now" conversation gently but firmly, because the client sees the mismatch themselves instead of taking our word for it. Seeing your own numbers not add up is a lot more persuasive than being told they might not.

We don't ask small clients to build enterprise-grade attribution systems before spending their first ad dollar. We ask for the minimum that lets us trust the direction of the numbers: one clearly defined conversion event, tested manually, with a rough baseline from existing traffic to compare against once ads go live.

That's usually a day or two of setup, not weeks, and it scales up as the ad budget itself scales up. The goal at small budgets isn't perfect attribution, it's trustworthy enough attribution that a decision to spend more or pull back is based on something real.

We'd rather lose a week of a client's patience than a month of their ad budget spent chasing numbers that were never real in the first place. Most clients, once they understand the tradeoff clearly, agree that's the right order to do things in.

This is also, frankly, a lesson we learned the expensive way ourselves on an early project where we skipped this step under time pressure, and spent weeks unwinding decisions built on numbers that turned out to be wrong. We don't skip it anymore, for anyone, regardless of how much they want to move fast.