Which figures already mean something in week 2, and which ones only after three months?

Most campaigns are not stopped too early because they are bad, but because they are judged on the wrong metric at the wrong time. Looking at the cost per customer in week two is like digging up a tree to see if it is taking root yet.

Below per phase which figures are already reliable and which are not yet.

Week 1: technique only

In the first week you do not assess any result. You check whether the machine is running.

• Is the campaign delivering? If hardly any impressions are coming in, something is wrong with the budget, targeting or ad approval.
• Is the measurement coming through? Run a test conversion yourself and check whether it appears in the advertising platform. This is the most important check of all, and the one most often skipped.
• Are the destinations correct? Does every advert end up on the correct page, and does that page load quickly enough.

What you do not do in week 1: adjust based on performance. The algorithm is in the learning phase and does not have a pattern yet.

Weeks 2 and 3: the first reliable signals

Now figures are coming in that actually mean something, because they are based on large volumes: you have thousands of impressions and clicks, but not yet of conversions.

• Cost-per-click. If that is well above what you expected, usually the target audience is too narrow or the competition in that auction is too high.
• Click-through rate. Indicates whether your advert matches who sees it. Low almost always means: wrong message for this stage, or wrong target audience.
• Post-click behaviour. Do people stay on the page or do they leave within a few seconds? That is an early signal regarding the quality of the traffic.

What you're not evaluating here yet: cost per conversion. With a handful of conversions, that figure changes dramatically due to chance.

Month 1: direction, no judgement

After four weeks you can see whether a campaign is moving in the right direction. The number of conversions starts to become stable enough to see differences between adverts or target audiences.

This is the moment for the first real optimisations: shifting budget between ad sets, turning off weak ads, and capitalising on the clearest audience differences. Still not the time for a final verdict on the channel.

Month 3: only now the real conversation

Most purchases are not impulse buys. Depending on the channel and your decision-making cycle, it takes two to six months for people to complete the customer journey. That means this month's revenue partly stems from advertisements from two months ago.

Only when at least one full decision cycle has passed can you look at the numbers that really matter:

• Cost per application, over enough applications to be reliable.
• Quality of those enquiries: how many become customers, and what are those customers worth.
• Whether the channel also helps the rest of your marketing, for example because more people come directly to your site.

The hidden cost: starting over

Every adjustment restarts a campaign's learning phase. A team that makes weekly adjustments out of impatience keeps the campaign permanently in the phase where it is most expensive and least predictable.

That is the paradox of judging too early: you want to see results sooner, and you ensure precisely that it takes longer.

Practically speaking: what do you agree on in advance?

Establish in advance what you will look at and when. That prevents both panic decisions and letting something run for too long that really isn't working.

• Week 1: MOT test, nobody looks at the result.
• Weeks 2 and 3: cost-per-click, click-through rate, post-click behaviour.
• Month 1: first optimisation round, determining direction.
• Month 3: the conversation about continuing, scaling up or stopping.

And agree on the outcome straight away. A campaign that doesn't pan out after three months is allowed to stop. That isn't a failure, that's the test you wanted to run.