~$ rebootworks
growth

before you scale the ad budget: five things that make paid traffic measurable

·13 min read·Dominick Brasileiro

why the conversion count in your ads dashboard is probably wrong, and what to fix before you spend another dollar.

a dashboard of campaign numbers being checked

Most paid ad accounts we audit are optimising towards a number that isn't real. The dashboard says forty conversions last month, but some of those fired twice, a share were spam form fills, a few were your own staff testing the page, and the landing page lost a chunk of the clicks you paid for before it finished loading on a phone. The platform doesn't know any of this, so it faithfully finds more people who look like the ones it counted.

Google and Meta are very good at spending money. Whether the spend does anything depends on five things that sit outside the ads platform, and most of them are technical. That's why they get skipped: the agency running the ads usually can't change the website, and the developer who can isn't looking at the ads.

How the money gets wasted

Three failures account for most of it, and they compound.

  1. The algorithm learns from what you tell it. Every modern bidding strategy optimises towards the conversion events you send. If those events are inflated, duplicated, or counting the wrong thing, the platform gets better and better at buying traffic that produces more of the wrong thing.
  2. The landing page wasn't built for the campaign. The ad promises one thing, the page opens with a hero about something else, offers six links and a slow carousel, and asks for eleven form fields. Half the visitors you paid for leave before the form exists.
  3. The account is structured how the platform wants rather than how you sell. Performance Max claims your brand searches and reports them as wins. Broad match spends on queries you would never approve. Nobody separated the people who already know you from the people who don't, so the numbers for both are blended into one flattering average.

None of this shows up as an error. The account runs, the report arrives, and the only symptom is that the leads don't turn into customers at the rate the numbers imply.

The five things to fix first

1. Verify the tracking before the first dollar is spent

Pick the one event that means money: a submitted enquiry, a booked call, a purchase. Then submit the form yourself and watch it land exactly once in GA4, once in Google Ads, and once in Meta. If it fires twice, the pixel is installed twice, usually once by a plugin and once by hand, and every conversion figure you have ever been shown is double.

Then move the tracking off the browser where you can. Meta's Conversions API and Google's enhanced conversions send the event from your server, so a blocked script or a privacy setting doesn't erase it. A meaningful share of real conversions never reach a browser-only pixel, and the platform optimises around the gap without telling you.

2. Count leads, not form fills

A form submission is a proxy. What you want the platform to optimise for is a lead someone on your team would actually call back. The fix is to send that judgment back to the ads platform as an offline conversion: mark the lead as qualified in your CRM or spreadsheet, and upload or sync that status against the click that produced it. Within a few weeks the bidding shifts from "people who fill in forms" to "people who turn out to be worth calling", which is the only outcome you were ever paying for.

3. Build the landing page for the campaign

One page per offer, opening with the same words as the ad, asking for one action, and nothing in the way of it: no main navigation, no unrelated links, a form short enough to finish standing up. Test it on a phone on mobile data, not on your office wifi. If it takes more than a couple of seconds to become usable, you are paying for clicks that leave before they arrive.

This is the step that stalls in most retainers. The ads team writes the brief, the developer queue is six weeks long, and the campaign runs against the homepage in the meantime. If the people running your ads can also build and swap the page, it happens the same week.

4. Structure the account how you sell

  • Separate brand from non-brand. Searches for your own name would have converted anyway; counting them as campaign wins makes everything else look cheaper than it is.
  • Keep Performance Max off your brand terms with exclusions, or it will spend its budget buying the traffic you were already getting for free.
  • Use broad match only once the conversion data is clean. Fed bad signals, it is the fastest way to spend on queries you never wanted.
  • On Meta, separate people who have already visited from people who haven't. They need different messages, and the cost to convert each is not comparable.

5. Own the accounts

The Google Ads account, the Meta Business account, the pixel, the GA4 property and the domain they all point at should be in your company's name, on your card, with whoever runs them added as a manager. An agency that owns your ad account owns your conversion history, your audiences and your negative keyword lists, which means leaving them costs you the year of learning you paid for. Ask for it in writing before the first campaign launches, and treat a refusal as the answer.

Reading the numbers

Once the five things above are in place, the report you want is short: spend, leads, cost per lead, and once the offline data is flowing, cost per qualified lead. Plus what was scaled, what was cut, and why, in writing, every week. A month of that is enough to decide whether to scale, fix, or stop. If the report can't support that decision, it isn't doing its job.

Resist reading it daily. Small budgets produce noisy numbers, and reacting to a bad Tuesday is how accounts end up restructured every fortnight and never learning anything.

When paid isn't the answer yet

Sometimes the honest recommendation is to wait. If the budget can't buy enough clicks to reach a statistically meaningful number of leads in a few weeks, you will spend the money and learn nothing. If nobody is going to call the leads back within a day, the ads are working and the business isn't. If there is no page worth sending traffic to, the page comes first.

In those cases the better use of the same money is usually the slower channel: fixing the site, publishing the pages buyers are searching for, and getting named in AI answers. Paid ads work on day one and stop the day you stop paying. Search compounds. The right setup is usually both, reported separately so you can tell them apart.

The short version

Verify the tracking before you spend. Feed qualified leads back, not form fills. Build the landing page for the campaign and make it fast on a phone. Structure the account so brand and non-brand are never blended. Own every account in your own name. Then raise the budget.


We run this as part of our growth practice: an account and tracking audit first, the technical fixes done by us the same week, then campaigns reported in cost per lead. Ask for an audit and we'll tell you whether your current numbers are real.