Q4 does not simply add traffic to an affiliate program. It changes the shape of the funnel. A discount resets the price every commission is calculated from, deal sites surface with your brand in their titles, and a slice of the orders you paid commission on gets refunded in January, after the money has already left your account.
The programs that come out ahead treat the burst as a project with a calendar. The ones that guess announce a code on the morning of, watch clicks triple, and spend February reconciling. Here is the version that holds up.
TL;DR
- Commission follows what the customer actually pays. A 30% discount cuts the commission base by 30%, so decide discount depth and commission rate in one conversation.
- One offer, one deadline, one code. Overlapping codes make attribution unreadable, and every extra code is another one that can leak.
- Brief your ten best affiliates six weeks out. Anyone with a real audience books Q4 early, and a late brief gets the leftovers.
- Refunds arrive in January. Hold commission on burst orders until the return rate settles instead of chasing money back.
- The burst is not the goal. A Q4 cohort is worth more in February, when you know who retained customers.
Why Q4 breaks a program that works in March
An evergreen program has three comfortable properties: the sale lands weeks after the click, the customer pays full price, and the commission is a predictable fraction of list. Q4 removes all three.
The discount rewrites the commission base. Every provider Sled integrates with reports the amount actually charged, after discounts and before tax: Polar's net amount, Paddle's pre-tax subtotal, Stripe's charged amount. Commission is a percentage of that number, so a deeper discount quietly reduces what the affiliate earns and the margin it comes out of. If you want an affiliate to earn the same dollars on a promotional order, the rate has to move with the discount.
Traffic quality falls as volume rises. Black Friday brings people hunting for any discount on anything. Judging an affiliate on burst-week conversion rate is how you fire a good partner by accident.
Calendars are already full. Publishers with real audiences planned holiday content in September. A brief sent the week before does not create a slot, it apologises for missing one.
The refund tail outlives the sale. Subscription refunds and chargebacks from a November promotion land across December and January, after you have recorded payouts.
The eight week calendar, week by week
Relative timing beats fixed dates, though it helps to know that Black Friday 2026 falls on November 27, with Cyber Monday on November 30. Count back from the day your offer goes live.
| When | Your job that week | What you ask affiliates for | What breaks if you skip it |
|---|---|---|---|
| Eight weeks out | Agree the offer, the discount depth and the window's commission rate | Nothing yet, this is internal | You improvise the rate under pressure |
| Six weeks out | Brief your ten best affiliates one to one | A yes or no, plus their format | They commit to someone else first |
| Four weeks out | Publish the offer, deadline and creative assets | Their intended publish date | Affiliates write their own copy and get the offer wrong |
| Two weeks out | Announce to the full list, publish the terms addendum | Which channel they will use | You find a channel conflict after launch |
| Launch week | Watch the webhook log daily, answer questions same day | Their live URL, so you can verify tracking | A broken link goes unnoticed for three days |
| The weekend itself | Watch for coupon-site hijacking and brand bidding | Nothing, do not bother them mid-push | You hear about hijacking from a customer |
| The week after | Second wave, only if the offer is genuinely ending | A short follow-up from the best performers | The tail of the campaign converts nobody |
| Two weeks after | Reconcile conversions against the export, hold payouts | Their own numbers, if they track them | Payout arguments start in January |
The two rows people skip are the first and the last: the meeting where rate and offer are decided, and the reconciliation two weeks after the dust settles.
Pick one offer and one deadline
Most of the value here is created by saying no to variants: one discount, one code, one window, one rate. Each extra variant costs attribution clarity, and every additional code is another one for a coupon site to leak.
- Tie the rate to the depth. If affiliates should earn the same on a discounted order, bump the rate for the window. Sled sets commission percentage per affiliate, overriding the program default, so a Q4 bump can apply to five named partners instead of everyone.
- Think about annual versus monthly. A discounted annual converts the burst into cash up front and a customer still around next November. A discounted monthly keeps the churn door open, which is the door a deal hunter walks through. Expect more refunds on monthly.
- Publish the closing time with a timezone. "Ends Sunday" produces Monday questions.
- Decide whether you will extend, and do not pretend otherwise. Affiliates compare notes, and a program that cries wolf on one deadline gets ignored on the next three.
- Say that earned commission is unaffected. An offer change should never rewrite commission on conversions that already landed.
What a discount does to commission math
This table is the argument for deciding depth and rate together. Assume a plan that lists at $99 per year, or $29 per month.
| List price | Discount | Customer pays | Commission at 30% | Commission at 40% | You keep at 30% |
|---|---|---|---|---|---|
| $99.00 | none | $99.00 | $29.70 | $39.60 | $69.30 |
| $99.00 | 20% | $79.20 | $23.76 | $31.68 | $55.44 |
| $99.00 | 30% | $69.30 | $20.79 | $27.72 | $48.51 |
| $99.00 | 40% | $59.40 | $17.82 | $23.76 | $41.58 |
| $99.00 | 50% | $49.50 | $14.85 | $19.80 | $34.65 |
| $29.00 | none | $29.00 | $8.70 | $11.60 | $20.30 |
| $29.00 | 30% | $20.30 | $6.09 | $8.12 | $14.21 |
| $29.00 | 50% | $14.50 | $4.35 | $5.80 | $10.15 |
Two things fall out. A 50% discount with a 40% commission on annual leaves you $29.70 of a $99 product, before support costs. That can be a fine trade for cash flow, but only if the number is a decision rather than a surprise. And the affiliate's payout moves with your discount even though their work did not change, which is why a rate bump is usually cheaper goodwill than a deeper discount: raising 30% to 40% on a discounted annual order costs $6.93 and makes them whole, while another 20% off the list costs you margin on every order, promotional or not.
If you pay recurring commission, the same logic runs forward, because the renewal is billed at the customer's post-promotion price. Recurring versus one-time commissions covers where each model breaks.
Brief your affiliates like a launch, not a newsletter
A code and a date produces a vague post about a vague deal. The affiliates who deliver know exactly what to say and have the assets to say it with.
- The offer in one sentence, written so it can be quoted verbatim, including what the product does for someone who has never heard of you.
- The deadline as a timestamp, plus what happens after it.
- Two or three example posts the affiliate can edit rather than write from scratch. Sled's Share Kit ships pre-written X and LinkedIn copy per program, and you can replace it in Settings under Share Kit post templates using
{link},{name}and{commission}placeholders. - Images they are allowed to post: checkout screenshots with the discount applied, a product GIF, a before and after. Link the files you already host once and every affiliate sees them as a gallery on their Share Kit page.
- A tracking link per affiliate, not a code they have to assemble. Test one yourself before the brief goes out and check the webhook log after your own test purchase.
Two emails to the list for the whole campaign is usually right: the brief, and one reminder in the final 48 hours. Email affiliates weekly through November and you are competing with the people who pay them.
The reminder sequence, and what already runs by itself
Three reminders are built into a modern program, which frees your Q4 attention for the fourth.
| Moment | What goes out | Who triggers it |
|---|---|---|
| An affiliate joins or is approved | Welcome email with their referral link and dashboard link | Automatic |
| Their first conversion lands | A first commission email, the one that gets people to post again | Automatic |
| They joined but never shared anything | A nudge with a direct link to their Share Kit | You, as a bulk action |
| The last 48 hours of your window | Whatever you write, to the people who said yes | You, by hand |
The bulk nudge is deliberately narrow: it targets affiliates with no shares and no conversions, skips anyone nudged in the last week, and caps the batch so the send completes. The cooldown matters more than it sounds, because an affiliate reminded four times in November will not promote you in December. For the manual one, the welcome email that gets opened works as a structure: here is your link, here is what to do with it, here is how long it takes.
Attribution traps that appear every November
Burst traffic finds every weakness in your setup at once. None of these are exotic, and each has a boring fix.
| Symptom | Usual cause | Fix before the burst |
|---|---|---|
| Clicks spike, sales do not move | Click farms and deal-site scraping | Judge on conversions, not clicks |
| A coupon site claims sales it did not drive | Last-click on a checkout someone else influenced | Review the affiliate list weekly, not monthly |
| Referrals vanish between click and purchase | Params stripped by an email client or in-app browser | Use the first-party tracking script |
| The same order appears twice | Webhook retries, delivery order is not guaranteed | One conversion per order id |
| Commission paid on a self-referral | Same person, two email addresses | Keep self-referral blocking on |
| A trial signup earns nothing, then pays later | The first order was $0 | Store the referral on the zero-value order |
The last-click problem deserves its own sentence, because a coupon site does not have to earn your sale to be credited for it. It only has to be the last link clicked before payment. Since cookie windows are measured in months by default, a deal site that appears in November can claim credit for a customer a newsletter affiliate introduced in September. The guide to attribution models and cookie windows covers the tradeoffs, and the fraud playbook covers the patterns that are deliberate.
Budget for the January refund wave
On a subscription product, a customer who bought in November to get the deal and cancelled in December is not an edge case, it is a percentage. If you have already paid the affiliate, recovering that commission is a conversation rather than a transaction.
Sled is built for the refund arriving in either order. A refunded order marks its conversion refunded and the commission drops out of the balance, including when the refund webhook arrives before the sale it refers to, in which case the conversion is created already marked refunded and no first commission email goes out for money that was clawed back. When the refund beats the sale walks through that race in detail.
What you control is the cash planning:
- Hold payouts on burst orders for at least one refund cycle, 14 or 30 days, and mention it in the brief so December is not a surprise.
- Use a minimum payout threshold rather than paying every small balance. A default of $20 per program keeps payment runs manageable and stops a clawback from creating a negative balance.
- Pay partially when the numbers are ambiguous. Paying less than the full owed amount settles the oldest commissions first and leaves the contested part pending, which is a better conversation than paying in full and asking for it back.
- Keep a list of refunded orders with their affiliate code, because that list answers "why did my balance drop" in January.
What to check every week during the burst
Fifteen minutes a week beats one long session in December, and it catches a broken tracking link while the campaign can still be saved.
- Clicks against conversions by affiliate. Thousands of clicks and no orders is either the wrong audience or a farmed link.
- Refunds attached to new orders. A refund rate that moves during the burst is a signal about the discount, not about the affiliate.
- The webhook log. Failed deliveries, ignored events and duplicates are listed with reasons instead of leaving you guessing why a sale did not appear.
- The conversions view, filtered to the window. It uses the same filters as the CSV export, which is what you hand to whoever does your books.
- One question to your agent. With a read-only API key connected, "which affiliates have clicks but no conversions since November 1" comes back with the numbers the dashboard shows.
If you do one thing only, check click to conversion on your top five affiliates. That is where the money is either made or quietly lost.
The January debrief
Do this in the second week of January, when refunds have mostly settled and nobody remembers the details.
- Segment by affiliate type. Newsletter, comparison site, coupon site, existing customer. Averages hide the fact that one group repaid its cost several times over and another did not.
- Measure cost per new customer, not cost per sale. Commission as a percentage of revenue flatters a deep discount. Compare dollars paid against customers still subscribed 60 days later.
- Reward the ones who stayed live. A few affiliates keep the offer page up or write a follow-up. Those are the people to give early access next year.
- Record offer depth and refund rate together. The pair is what tells you whether last year's discount was too deep, and it is the only honest way to answer the question next October.
- Ask existing customers to promote it. A large share of Q4 wins come from people who were already recommending you, and turning customers into affiliates is a two-week job. Do it in October rather than November.
For the longer version of that measurement argument, including payback months and incrementality rather than one campaign's cost per customer, affiliate program ROI has the cohort math.
FAQ
Should I raise the commission rate or deepen the discount for Black Friday? Raise the rate for the window. The discount costs margin on every order, including ones that would have converted anyway, while a temporary rate bump costs you only on orders an affiliate actually drives.
Do affiliates get paid on discounted orders? Yes, on the discounted amount. Commission is calculated from what the customer actually paid, after discount and before tax, so a 30% discount reduces both what you collect and what the affiliate earns. If they should earn the same on a promotional order, adjust the rate rather than assuming they will accept less.
How long should a Black Friday affiliate window run? Long enough to be planned and short enough to be believable. Two to four weeks ending on Cyber Monday is the common shape, but the affiliates doing the real work need the offer four to six weeks before it starts.
What if an affiliate leaks my discount code to a coupon site? Decide the policy before it happens rather than during the argument. Options range from ignoring it, to excluding coupon and deal-site traffic from commissions, to removing the affiliate. Whatever you choose belongs in the program terms, because a rule applied after commission has been earned reads as a retroactive change, and those do not survive a public argument.
If your burst is going to be judged on attribution anyway, read the cookie window and attribution guide next, or start a program that tracks orders, refunds and payouts in one place.