The majority of businesses that think their marketing isn't working actually have a measurement problem, not a marketing problem. Without reliable measurement, you can neither cut what doesn't work nor strengthen what works, because you simply don't know which is which. The correct order of operations is to fix the metric first, and only then adjust the ad spend.
The most common, misdiagnosed problem
A business spends on advertising for months, sees disappointing results and concludes that advertising does not work in its sector. In a surprising number of cases, the campaigns are not the main problem. Measurement is unreliable, so no one can tell what is actually happening.
The order of operations that changes everything
Repair the measure first. A clean analysis, with real conversions clearly defined for your company, not generic definitions that mean nothing concretely.
Follow conversions from end to end. From the first click to the application received, via the completed form.
Prepare dedicated pages before spending. An advertising campaign deserves better than a simple link to the general home page of the site.
Spend only once these three steps have been completed. At this point, every dollar spent produces actionable data to adjust what’s next.
Spending before completing these steps is buying traffic that you will never be able to properly judge.
| Angle | Dimension to examine | Use in the decision |
|---|---|---|
| 01 | The most common, misdiagnosed problem | Setting the Context |
| 02 | The order of operations that changes everything | Identify dependencies |
| 03 | The most common errors I encounter | Compare options |
| 04 | What attribution can do, and what it can't do | Preparing for the next step |
The most common errors I encounter
Your own team's internal traffic is never filtered, artificially inflating all numbers. No conversions are actually defined, or the defined conversions don't mean anything concrete to your business, such as simply having visited a page. The forms aren't tracked at all, so a request could come in without analytics ever recording it. All campaigns point to the same general home page rather than dedicated pages for each specific offer. No one ever compares the advertising platform's numbers with the site's analytics, even though the two should match.
Set your decision
Select the dimensions that describe your situation. The result is to organize the next conversation, not to replace an analysis.
Choose the dimensions that apply to show the recommended level of attention.
What attribution can do, and what it can't do
A well-configured measure can show you which campaigns usually precede the applications received, compare different periods between them, and reveal which channels actually produce nothing concrete. It cannot perfectly reconstruct the entire journey of a customer who saw an advertisement on his phone, searched your name three days later on a computer, then called from a third different device. Current privacy restrictions also reduce the visibility available. An honest and imperfect measure is infinitely better than a measure that claims to be perfect when it is not.
The five elements of a complete infrastructure
A well-configured analysis to know what is actually happening on the site. A tag manager to install and control the tools without risking breaking the site with each addition. A conversion tracking that links each application received to its source. Ad pixels properly installed to allow remarketing and campaign optimization. Landing pages dedicated to each specific offer rather than a single generic page for everything.
What I see on the ground
When a company tells me that its advertising does not work, my first reaction is always to check the conversions before looking at the campaigns themselves. In a striking proportion of cases, advertising actually worked properly. It was the measure that was broken, and the company cut off the wrong budget thinking it would solve the problem.
When this guide does not apply
If you do not make any paid advertising and your traffic comes only from organic sources or word of mouth, some elements of this guide are less priority for you immediately. However, the basic measure is still useful in understanding the behaviour of your visitors, regardless of their origin.
Frequently asked questions
01How long does it take to collect reliable data after tracking is repaired?
The configuration itself is done quickly, but truly actionable trends generally require a few weeks of data collection.
02Do we need a minimum advertising budget for measurement to be worthwhile?
The basic measure is useful regardless of the budget. To obtain statistically reliable data on the campaigns themselves, a reasonable minimum budget becomes necessary.
03How can we tell whether our current marketing numbers are reliable?
The only way to know is to test concretely, for example by filling out your own contact form and checking whether a conversion appears correctly in your data.
In-depth guide prepared by GXN based on real business situations. Recommendations remain independent of brands and providers.
References to verify for your situation
These references support the external rules and frameworks cited in this guide. They do not replace legal or professional advice tailored to your organization.