Cookie windows are shrinking: what still works

A cookie window is the period after someone clicks your link during which a purchase still counts as yours. Thirty days used to be unremarkable. It is now the generous end of the range, and 24 hours is no longer unusual.

Sephora runs a 24-hour window. Steve Madden runs seven days. Charlotte Tilbury runs thirty. Those are three different businesses, and they demand three different approaches from you.

Program	Window	Rate	Avg order
Sephora	24 hours	Up to 5%	$120
Steve Madden	7 days	5%	$95
Charlotte Tilbury	30 days	Up to 8%	$105

Why advertisers shortened them

Not to be difficult. Two things changed.

Attribution got better, and advertisers could finally see how much affiliate commission was being paid on sales that would have happened anyway, someone already at the checkout who opened a new tab to find a voucher code. A shorter window is the bluntest available correction for that.

The second reason is that paid social got expensive, so more brands run their own retargeting. When a brand is already chasing the customer with its own ads, it is reluctant to also pay commission on a thirty-day-old click.

Where the short windows actually are

Shrinking is the trend. It is not yet the norm, and it is worth knowing which programs it has reached, because the answer is not the one most creators assume.

Of the 229 programs in our directory, 18 run a window of seven days or less and two of those are effectively same-session, at 24 hours. Another 27 sit between eight and fourteen days. The remaining three quarters are at fifteen days or more, and thirty is still by a wide margin the most common single figure.

Fashion is the only category whose median window is below thirty days. It sits at 21, and seven of the 32 fashion programs we list are at seven days or under. Every other vertical in the directory has a median of exactly thirty, including the ones where the purchase takes longest to decide.

That last point is the uncomfortable one, because a short window is not distributed according to how quickly people buy. It is distributed according to how much retargeting the brand does.

Expedia$620

7-day window, 3% per booking

Hotels.com$420

7-day window, 4% per completed stay

Pottery Barn$340

7-day window, 3% per sale

Wayfair$260

7-day window, up to 7% per sale

Best Buy$210

24-hour window, 1% per sale

J.Crew$145

7-day window, 5% per sale

Banana Republic$130

7-day window, up to 5% per sale

Sephora$120

24-hour window, up to 5% per sale

Every program in the directory with a window of seven days or less, ranked by average order value. Compiled from the published figures on each program page. The top four are furniture and travel, categories nobody buys in a single session.

Read that chart as a list of the places where the window is most likely to cost you money. A $620 flight booking on a seven-day window and a $340 sofa on the same window are not short-window purchases. They are long deliberations being asked to behave like impulse buys, and the gap between the two is revenue that quietly goes to whoever the buyer clicks last.

What a short window actually breaks

The format that suffers is the evergreen roundup: the "best running shoes" post that accumulates traffic for eighteen months.

It still gets clicks. It just stops converting them, because the reader who arrives in March, saves the page, and buys in April is outside the window. Your analytics show healthy traffic and your commission report shows very little, and the two never reconcile because nothing in the click data tells you the sale happened.

The same applies to a video that keeps surfacing months after posting. Reach without attribution earns nothing.

What still works under a short window

Content published close to the moment of intent. A sale-day post, a launch review, a "this is 30% off today" story. If the purchase decision and the click happen in the same session, the window length is irrelevant.

Specific product recommendations rather than category overviews. "Here is the exact bag I use" converts within the session far more reliably than "the ten best bags", because the reader has one decision to make instead of ten.

Formats with a built-in deadline. Sales, restocks, limited colorways. Urgency compresses the gap between click and purchase, which is exactly what a short window rewards.

What to do about the long-window programs

Do the opposite. A thirty-day window is what makes a back catalogue worth building.

If you are choosing between two programs and one runs thirty days, that program is where your evergreen content should point. Buying guides, comparison posts and anything you expect to keep ranking should be aimed at advertisers whose window is long enough for the traffic to still be worth something in week three.

This is a content-planning decision more than a program-selection one. Most creators pick the program first and then write whatever they were going to write anyway. The order should be reversed.

Eight programs in the directory run longer than thirty days, and seven of those are at 45. Abercrombie & Fitch is the one worth knowing about if you make fashion content, because it is six and a half weeks in a category whose median is three, and it is the only place in fashion where a back catalogue behaves the way a back catalogue is supposed to.

The window is not the same clock as the hold period

These get confused constantly, and confusing them produces a specific wrong conclusion, so it is worth pulling apart.

The cookie window decides whether you are credited. It runs from the click and it closes when it closes.

The hold period decides when you are paid. It runs from the sale, it is typically 30 to 60 days in retail and 60 to 90 in travel, and it exists so returns and cancelations can land before the money is released.

They are unrelated settings. A 24-hour program can hold commission for two months, and a 45-day program can validate quickly. What matters practically is that the two failure modes look identical from where you are sitting. A sale you never got credited for and a sale that is sitting in hold both show up as clicks with no money attached, and only one of them is ever going to resolve.

The way to tell them apart is time. Wait one full validation cycle. What appears is a hold-period question, and what never appears was a window question. Guessing in week two, which is when everybody wants an answer, gets it wrong roughly half the time. The full validation timeline is worth reading before you conclude anything about a program from its first fortnight.

The practical check

Before you build content around a program, look at the window and ask one question: how long between someone seeing this and someone buying?

If the answer is minutes, any window works. If it is a fortnight, a $400 purchase, a considered replacement, anything needing a conversation with someone else first, a seven-day window will lose you most of the revenue and you will never see it happen.

Every program page in the directory publishes the window, and the long cookie window list collects the ones at the generous end.

One thing worth not doing

Do not choose programs on window length alone. A thirty-day window on a program your audience has no interest in is worth nothing, and a 24-hour window on a brand they already buy weekly can be excellent.

Window length tells you which format to use. It does not tell you which brand to promote.