Affiliate Cookie Duration Statistics: What Programs Offer in 2026
The real distribution of cookie windows across 134 programs with stated terms, and how to judge whether a window is long enough for your audience.
Published July 17, 2026
The median affiliate cookie window in our directory is 45 days. 45% of programs sit exactly at 30 days, 42% offer 31 to 90 days, and 12% offer more than 90 days. 6 programs advertise lifetime attribution. These numbers are computed from the live dataset of 160 programs every time this page builds; 26 listings state no usable window and are excluded rather than guessed.
The distribution
Cookie windows across every program in the directory that states one:
| Cookie window | Programs | Share |
|---|---|---|
| Under 30 days | 2 | 1% |
| 30 days (the default) | 60 | 45% |
| 31 to 60 days | 30 | 22% |
| 61 to 90 days | 26 | 19% |
| 91 to 180 days | 8 | 6% |
| Over 180 days | 8 | 6% |
The longest stated windows in the directory right now:
| Program | Cookie | Category |
|---|---|---|
| DreamHost | lifetime | Hosting |
| MailerLite | lifetime | Email Marketing |
| Mangools | lifetime | SEO Tools |
| Moosend | lifetime | Email Marketing |
| Systeme.io | Lifetime | SaaS |
At the other end of the scale sit the sub-day windows. The most famous is Amazon Associates at 24 hours, and travel booking programs often use session-only attribution: the visitor books in that sitting or you earn nothing. Those short windows are not stinginess for its own sake; they reflect how those products are bought, which is the key to reading this whole table.
Cookie windows by category
The window a program offers correlates with its category, because the window is really a statement about how long the merchant believes their buyers deliberate. Median stated window per category, computed from the live data:
| Category | Programs with stated windows | Median window |
|---|---|---|
| SaaS | 19 | 60 days |
| AI Tools | 17 | 60 days |
| SEO Tools | 10 | 60 days |
| Email Marketing | 9 | 120 days |
| Hosting | 13 | 60 days |
| VPN | 11 | 30 days |
| Education | 8 | 30 days |
| Finance | 12 | 30 days |
| Design | 4 | 30 days |
| Ecommerce | 7 | 30 days |
Two readings of this table are useful. Within a category, a program offering meaningfully more than the category median is paying for conversions its competitors let expire, which is a real edge when you promote in that category. And across categories, the medians tell you where window length is a live negotiating dimension (B2B-leaning software) versus where it is a fixed convention nobody competes on.
What a cookie window mechanically is
When a visitor clicks your affiliate link, the program (or its network) drops an identifier in the visitor's browser and usually logs the click server side with a click ID. That identifier says: this person arrived through this affiliate. If the visitor completes a purchase before the identifier expires, the sale is attributed to you and the commission enters your pending balance. After expiry, the same purchase pays you nothing.
Three details decide how that plays out in practice:
- Last click wins. Nearly all programs overwrite the stored referrer when the visitor clicks another affiliate's link. You can do the persuading and lose the commission to a coupon site the buyer visits at checkout. A few programs run first-click attribution, which inverts the game; the terms page will say.
- Browsers fight cookies now. Safari's tracking prevention caps many cookie lifetimes at days, not months, and content blockers strip tracking requests entirely. A stated 90-day window is an upper bound. Serious programs supplement with server-side click IDs and first-party attribution so the referral survives the browser; this is a real reason to prefer programs on modern tracking infrastructure.
- Account-based attribution beats cookies. When a program credits whoever referred the account at signup (the model behind the lifetime windows above), cleared cookies and device switches stop mattering. For products with free tiers where signup happens fast but payment happens months later, this is worth more than any cookie length.
Why 30 days became the default
No standards body picked 30 days. It won because it is the schelling point between two pressures: long enough to cover the deliberation cycle of a typical software purchase (research, trial, one billing thought), short enough that merchants are not paying commissions on sales their own remarketing would have closed anyway. Once the biggest networks shipped 30 days as the default field value, thousands of programs simply never changed it.
That means a 30-day cookie tells you nothing about a program except that it took the default. The information is in the deviations. A program that chose 90 or 120 days is signaling it wants affiliates enough to pay for conversions it might have gotten free, the way Frase (120 days) and Leadpages (90 days) do. A program that chose 7 days is telling you it only values the click-to-checkout handoff.
When a short cookie is fine, and when it kills the deal
Match the window to the deliberation time of the purchase, not to a universal standard:
- Impulse and low-ticket (under $30, consumer): conversion happens in hours. A 24-hour window costs you little; Amazon built the biggest affiliate program on earth on one.
- Standard SaaS ($10 to $100/mo): a visitor trials, thinks, and converts within a week or three. 30 days covers most of the curve; 60 to 90 days picks up the long tail.
- High-ticket and B2B ($100+/mo, team decisions): buying cycles run 30 to 90 days by themselves. A 30-day cookie can quietly forfeit a large share of your real referrals, and window length becomes a first-order factor in choosing between competing programs.
- Free-tier products that upgrade later: the cookie needs to survive until payment, or the program needs account-based attribution. This is where lifetime models earn their keep.
To see what a window is worth in your own numbers, our cookie window calculator models how much attributed revenue different durations capture given your audience's conversion lag. It makes the trade concrete in about a minute.
Attribution models: last click is the default, not the law
The cookie window says how long attribution lasts; the attribution model says who wins when several affiliates touched the same buyer. Knowing the model changes which content is worth producing.
Last click is the overwhelming default: whoever owns the most recent click before purchase takes the whole commission. It structurally favors bottom-of-funnel content (comparisons, reviews, deal pages) and punishes top-of-funnel education, because the person who taught the buyer loses to the person who caught them at checkout.
First click credits the affiliate who introduced the buyer, however long ago within the window. Rare, but where it exists it inverts the incentive: tutorials and discovery content become the winning formats. Programs advertise it when they have it, because it is a recruiting advantage with content publishers.
Coupon and code attribution bypasses clicks entirely: the buyer enters your code at checkout and the sale is yours regardless of cookies. It survives every browser restriction and every competing click, which is why creator-focused programs lean on it. If a program offers you a code alongside a link, the code is the more durable asset.
Account-based (lifetime) attribution ties the referral to the account created after your click, as covered above. The practical consequence for content: your job becomes getting the free signup, not the purchase, which suits products with generous free tiers.
Matching content type to window length
The window you need depends on what your content does to the buyer, and you can chart it fairly precisely:
- Deal and coupon content: the visitor arrives ready to buy; any window works, including 24 hours. Rate and clawback terms matter far more than duration.
- Product reviews and comparisons: visitors convert over days, not minutes. 30 days captures most of the curve; below 14 days you are donating conversions.
- Tutorials and how-to content: readers often act on the recommendation weeks later, when the problem recurs. 60 to 90 days meaningfully outperforms 30 here, and this is the content type where checking the window before choosing between rival programs pays most.
- Audience channels (newsletters, YouTube): subscribers click when the email lands and buy when the need arrives, which can be months apart. Prefer account-based attribution or 90+ day windows, and prefer codes wherever they are offered.
Cookie length interacts with commission size
Never read the cookie column alone. A program paying 40% behind a 24-hour window and a program paying 25% behind 90 days can be the same effective rate once you account for how many of your conversions each window actually captures. The math is simple: effective rate = stated rate × share of your conversions that happen inside the window. For a high-deliberation audience, the second program wins comfortably; for an impulse audience, the first does.
This is also why headline-rate comparisons between competing programs mislead. Rate ranges by category are in our commission rates data, and the recurring dimension (which dwarfs both rate and cookie over a year) is worked through in recurring vs one-time commissions.
Questions to ask beyond the number
- Does the window reset on a second click? Most programs restart the clock every time the visitor clicks your link again. A few do not.
- Does it cover the cart or the account? Some programs attribute only the first purchase; others credit everything the referred account buys within the window.
- What happens across devices? A cookie set on a phone does not exist on the buyer's laptop. Programs with logged-in ecosystems or server-side click IDs bridge devices; pure cookie programs do not.
- Is there a conversion deadline separate from the cookie? Lifetime-attribution programs sometimes require the referred account to become a paying customer within some period for the tie to hold.
Methodology
Every count and share on this page is computed at build time from the affiliatejob directory: 160 programs total, of which 134 state a concrete cookie window and enter the distribution. The 26 programs with no stated or applicable window (some finance and CPA programs attribute per action without a browser window at all) are excluded rather than estimated. Windows are as programs state them; browser privacy features can shorten real-world persistence below any stated figure, and we make no claim about effective in-browser lifetimes. The dataset is public at /data/programs.json, and changes to any program's cookie terms appear on the changelog.