Why Black Friday should start early

In late September I planned Black Friday with the team at a hand-dyed yarn brand, the small-batch studio from the acquisition case. Its regulars reorder every month and brought in 79% of revenue (share of store revenue) in the six months to July, so any sale that changes when they buy is a big decision.

We settled on two windows. Early Black Friday runs Friday November 6 to Monday November 9: a flat 20% off everything for the email list, not stackable, plus a fixed $5 credit by text for SMS subscribers that does stack. Black Friday itself runs November 27 to 30, through Cyber Monday, with chosen items on sale at different depths and no sitewide code.

This issue is the plan and the reasons for it. The results can wait until the quarter's orders are in.

Only three sales can be predictable

Earlier that month we agreed a rule that now governs every promotion this brand runs. Only three sales a year may be predictable: Memorial Day, Black Friday and Cyber Monday. Customers already plan around those. Everything else has to feel random.

A sale is predictable when a regular can tell it's coming, the same week every month or the same weekend every year. The cost is the full-price orders that wait for it. Someone who knows a code is due in two weeks holds the order, and what they think the yarn is worth drifts toward the sale price. The team had seen it in their own customers after repeated promotion cycles, down to a note of entitlement in customer service. As one of them said, "We trained them for this."

What should be predictable is the newsletter. This brand sends on Friday and Monday, and I want the list told that those are the days to expect email. For the rest of the year the tools leave the price alone: community make-alongs (this year's did very well with no discount), a double-points weekend that rewards the next order, and an occasional surprise drop.

What the early window protects

The early window sits inside the Black Friday slot, so there are still three predictable sales a year. Splitting the slot protects a few things.

First, the regulars. The 20% goes only to the email list, and the code isn't advertised. Last year it ran for two days through a separate double opt-in list of 2,273 profiles. This year it goes to the whole list, because every extra opt-in brings a wave of "Am I opted in?" emails to customer service. The emails invite people to heart what they want ahead of time, so their list is waiting on opening morning.

Then the inbox. By the last week of November a customer hears from nearly every brand they've bought from, most with a discount. In the first week of November there's far less of that, so the email with the year's biggest code has room to be read.

And the margin on the main weekend. Black Friday itself runs at a smaller discount, on chosen items, with reward points usable on top. Its new reason to shop is one limited colorway. Last year two weekend colorways underperformed, and the team's read was that Christmas palettes don't land with these customers and that splitting attention across two may have hurt. Autumn palettes have worked.

Keeping the code on the list

One shared code has a weak point. In October the 20% code was set up early with the Shop app sales channel switched on, the app showed it to shoppers, and 16 people used it before the window opened, several of them first-time customers.

That turned out to be useful. It showed the Shop channel brings in new customers, and the team is now open to Shop ads with a first-order coupon. For the early window, the code stays off every public channel, and since one code sent to a whole list tends to reach coupon sites, I'd use one-time codes per subscriber.

The one test in the plan

The 20% is last year's offer, which the team calls the best coupon they've run. The new piece is the $5 by text. Every email asks readers who don't get texts yet to add their number. SMS here is double opt-in, so nobody gets the $5 until they reply Y, and the emails remind them.

I pushed for SMS because the rules around email keep shifting, and a brand that reaches customers only one way is exposed. Email stays the main channel because it's the cheapest, with SMS as the fast backup.

The codes were tested together before anyone wrote copy. A non-stackable 20% and a stackable $5 should combine, but it had gone wrong the last time the team tried. They tested it live on a call in October, and the codes combined.

It isn't a clean experiment, since dropping last year's sign-up list also changed who gets the 20%. The SMS count can still be read on its own: how many email readers added a number between the first announcement and the last day, and how many still subscribe and order by spring.

Do this Monday

This takes an afternoon with last year's order export and a blank calendar.

1. List every discount you ran in the last 12 months, with dates and depth. Mark each one a regular could have seen coming.
2. Choose your predictable sales, three at most, and put them on the calendar first. Move everything else to dates customers can't guess, or swap it for something that leaves the price alone, like a make-along or bonus points on the next order.
3. Give your list its biggest code a few weeks ahead of Black Friday, for about four days, by email only. Check it isn't switched on for any public sales channel, and decide on one-time codes.
4. Plan the main weekend around chosen items and something new, so it doesn't need a sitewide code.
5. Ask whoever packs your orders how big a weekend they can absorb. This studio can take a big weekend, but anything larger costs them about a week, so nothing else large sits near Black Friday.
6. Pick one test and write down now what you'll count to judge it.

---

In one line: give your list the biggest code of the year in a short, quiet window a few weeks before Black Friday, and keep every other discount hard to predict.