The sale that shows up two weeks late

Here is an assumption hiding inside almost every campaign review: that a sale belongs to the week it happened in. You ran something on Monday, you look at the numbers on Friday, and whatever the numbers say is the verdict. It feels like common sense. It is also wrong for most things people buy with any thought at all, because the person who clicked on Monday is very often the person who pays two weeks later, after a second visit, a comparison with two alternatives and a conversation with whoever shares the card.

The gap between first contact and purchase has a boring name, conversion lag, and it quietly decides more campaign verdicts than creative, targeting or budget combined. Not because anyone argues about it, but because nobody measures it, so every review is done as if it were zero.

What the lag does to a Friday review

Invent some numbers, because the shape is the point. Say a campaign runs for one week and, in the end, earns twenty customers. Say your buyers behave the way many considered purchases do: a third of them pay within a few days of the first click, another third within two weeks, and the last third somewhere in the month after that.

On the Friday of launch week you see maybe five or six of those twenty. At that moment the campaign looks like it produced a quarter of what it will eventually produce. If you divide spend by what you can see, your cost per customer looks three to four times worse than it really is. That is the number people use to turn campaigns off.

Now run the same arithmetic in the other direction. The campaign you paused on Friday keeps "earning" for another month, because the clicks were already bought and the people are already deciding. Sales trickle in with nothing new running, and it becomes very tempting to credit them to whatever you launched next. One decision, made on a lagged number, and you now have two wrong conclusions: the first campaign looked bad, and the second looks better than it is.

the week you launch is rarely the week you get paid.

Find your own number before you need it

The lag is not a universal constant; it depends on the price, the category and how much of a stranger the buyer was. A cheap impulse purchase can close in minutes. A tool someone will pay for monthly, and has to explain to a colleague, often takes weeks. The only lag worth using is yours, and you can estimate it without anything fancy.

Take your last twenty or thirty customers and, for each one, find two dates: the first time they showed up (first visit, first signup, first email opened, whatever your tools can see) and the day they paid. Sort the gaps. The middle one is your typical lag; the one about three quarters of the way down is the one that matters for reviews, because it tells you when most of the results from a given week have actually arrived. If that number is nineteen days, then a campaign cannot be judged in fewer than nineteen days, no matter how uncomfortable the wait is.

If you cannot connect the two dates at all, that is useful information too. It usually means the first visit and the purchase are being recorded in different tools with no shared thread, which is the problem we walked through in three tools, three numbers, one signup. Fixing that is worth doing before any campaign review, because every lag estimate depends on it.

What changes once you know it

The first thing that changes is the calendar. Instead of reviewing on a fixed day, you review each campaign when its window has closed: launched on the third, lag of about three weeks, judged around the twenty-fifth. Anything you look at earlier is a progress report, not a verdict, and it helps to call it that out loud so nobody treats it as one.

The second thing is what you watch in the meantime. You do not need to fly blind for three weeks. Pick the earliest signal that reliably comes before a sale in your business, whether that is a signup, a trial started or a pricing page visit, and watch that during the wait. It will not tell you revenue, but it will tell you quickly if a campaign is producing nobody at all, which is the one case where waiting really is a waste. This is the same distinction as in when to kill a campaign, and when it just needs a week: an early number can rule a campaign out, it rarely rules one in.

The third change is quieter. When you look at a month of results, sales from the first half of the month tell you about that month's campaigns; sales from the last week or so are mostly still telling you about the month before. Mixing them is how a good month gets blamed for a bad one.

The tradeoff you are accepting

Waiting has a real cost. A campaign that is genuinely bad keeps spending for the extra two weeks you gave it, and that money is gone. The honest way to manage that is not to shorten the window but to shrink the stakes during it: run new things at a budget you would not mind losing for the length of your lag, and only raise it after a full window has closed with a verdict you can believe. Small and patient beats big and hasty, because big and hasty makes you decide on a quarter of the evidence.

The rule of thumb, then, fits in one line. Measure how long your customers take to decide, write that number down where you do your reviews, and do not let any campaign be judged before it has had that long to be right.