The attribution window and your sales cycle

The platform credits what closes inside its window. If your cycle is longer, part of the result disappears from the dashboard — and measuring how much is a single query.

Every ad platform credits conversions inside a window: seven days after the click, one day after the view, in Meta's default. It is a reasonable and arbitrary choice — and it only becomes a problem when your sales cycle does not fit inside it.

The useful question is not whether the window is fair. It is how much of your result falls outside it, and that is measurable.

Measuring your own cycle

Take the sales from the last period and, for each one, the distance in days back to the first contact or the originating click. The distribution answers everything.

In a business we measured, from first sales contact to purchase: 44.2% the same day, 19.5% within one day, 15.6% two to three days, 13% four to seven, and 7.8% between eight and fourteen days. Adding up, 92.2% closed within seven days — and around 8% escaped.

In the same business, measured from the ad click to the purchase, the distribution was wider: 32.8% the same day, 23% in one day, 21.3% in two to three, 9.8% four to seven, and 13.1% at eight days or more, with cases at twenty-five and thirty-one days. An average of 4.1 days.

Both distributions come from the same business and tell different stories. The distance from the click is always longer than the distance from contact — and it is the click distance the platform's window measures.

The caveat that prevents a wrong conclusion

There is a bias that fools almost everyone on the first measurement: you can only observe what has had time to happen. If your history starts three weeks ago, a sale that would take forty days has not happened yet, and therefore is not in the count.

That means the percentage you measure as "outside the window" is a floor, not a ceiling. With short history, it always looks smaller than it is.

What the window actually decides

It is worth understanding what changes when a sale falls outside. It does not disappear from your revenue — it disappears from that campaign's credit. The money arrived; the origin information did not reach the platform in time for it to learn from it.

The practical consequence is that cost per sale shown in the dashboard is always pessimistic for long-cycle businesses, and so is return. Cutting budget based on that number means cutting on an incomplete measurement, and the cut then justifies itself afterwards — less budget, fewer sales, worse number.

A detail that changes how you read last click

There is a common pattern in businesses with a sales team: the ad brings the person in, a rep talks over chat and sends the payment link. That link does not pass through the pages carrying campaign parameters, so it creates no new touch.

The result is that credit stays on an older click, sometimes weeks old. That is not a measurement error: it is the nature of last click. It describes where the person entered, not who closed the sale.

Confusing the two leads to strange decisions, like cutting the campaign that brings people because "the reps are the ones closing" — when it was the campaign that put the person into the conversation.

What to do with the number

• Measure your distribution before arguing about the window. Without it, the conversation is opinion.
• Compare platform credit with your books over the same period, and look at the size of the gap, not just the return.
• If a good share closes outside seven days, consider optimising for an earlier event in the funnel — conversation started, for instance — instead of waiting for the purchase to fit the window.
• When reading the dashboard, remember it measures origin, not closing merit.

In CrazyLeads the distance between first touch and sale is a query over the person's own history — and the report shows both what the platform credited and what your records say, side by side.