Dev.to WebDev πŸ›  Dev πŸ‘ 0 πŸ“– 3 min read

Your Black Friday posts need a kill switch, not a better caption

Black Friday is next Friday (Nov 27). Somewhere right now a person is lining up 30 scheduled posts for that weekend and feeling very organised about it. I know because I usually am that person. The captions are rarely t

Black Friday is next Friday (Nov 27). Somewhere right now a person is lining up 30 scheduled posts for that weekend and feeling very organised about it. I know because I usually am that person.

The captions are rarely the problem. The problem is what happens when reality changes at 11am on Friday and the queue doesnt know. Stock runs out, the checkout page goes down, the discount code breaks, a big news event makes your "SMASH THAT BUY BUTTON" post look tone deaf. And your scheduler keeps firing like nothing happened, every 2 hours, on 4 networks.

Quick disclosure: I work at ContentStudio (a social scheduling tool), so I see a lot of these weekends from the support side. None of this needs our product though. It works the same with n8n, a Google Sheet plus Apps Script, Zapier, or whatever homemade cron you have running on a VPS.

1. One flag that every scheduled promo checks

The simplest version is a single value somewhere boring. A cell in a Google Sheet, a key in Redis, an env var, a row in Airtable. Call it PROMO_STATUS with three states:

  • live means post normally
  • hold means dont post, keep the item in the queue
  • soldout means swap to the sold out variant (if you wrote one) or skip

Then whatever publishes your posts reads the flag right before sending, not when the post was created. That part matters. Checking at schedule time is useless, the whole point is that the world changed after you scheduled it.

In n8n that is literally a Google Sheets read node and an IF node before the publish step. In code its something like:

async function shouldPublish(post) {
  const status = await getPromoStatus(); // sheet cell, kv, whatever
  if (!post.tags.includes("bf-promo")) return { ok: true };
  if (status === "hold") return { ok: false, reason: "promo on hold" };
  if (status === "soldout") return { ok: true, swapTo: post.soldOutVariant ?? null };
  return { ok: true };
}

Only tag the promo posts. Your normal "here's a tip" posts can keep going, nobody needs those paused because the 40% code broke.

2. Make the flag flippable by someone who isnt you

If flipping it needs a deploy, or SSH, or you, it wont get flipped in time. Put it where the person watching sales on Friday can change it from their phone. A Sheet cell works surprisingly well for this. So does a tiny Slack slash command if you already have a bot.

Write down who owns it for each block of the weekend. Friday 9-14 is Anna, Friday 14-20 is Marco, etc. Sounds like overkill until its Saturday 8am and everyone assumes someone else is watching.

3. Let a health check flip it for you

You probably already have uptime monitoring (UptimeRobot, Better Stack, a cron that curls your checkout). Point one alert at the flag too. Checkout returns 5xx for 3 checks in a row, set PROMO_STATUS=hold automatically and ping a human.

Sending traffic to a broken checkout is worse than sending none. Its paid attention you're burning.

Same idea for the discount code. A tiny script that applies the code to a test cart every 30 min and checks the price actually dropped. Codes break more often than you'd think on BF weekend, usually because someone "just tweaked" a rule.

4. Log every skip

When the flag stops a post, log it with the post id, time and reason. On Monday you want to know which posts never went out so you can decide if they get rescheduled or deleted. Otherwise you find them in the queue in March, still promising 40% off.

2026-11-27T11:42:10+01:00  post=bf-ig-07  skipped  reason="promo on hold"

A Sheet tab works fine for this too. Doesnt need to be fancy.

5. Test the switch before Friday

Pick one dummy post scheduled 10 min from now, flip to hold, confirm it didnt go out. Flip to soldout, confirm the variant went. 15 minutes of testing on a Tuesday is a lot cheaper than finding out the IF node was reading the wrong column at peak.

The short version

  • one promo flag, read at send time, not schedule time
  • only promo posts check it
  • someone besides you can flip it from a phone
  • monitoring can flip it to hold
  • skips get logged
  • you tested it once

None of this is clever. Its just the boring thing that would have saved a few people I know a very bad Black Friday afternoon. Good luck next week, and please write the sold out caption now while you're still calm.

πŸ“° Read the original article on Dev.to WebDev

Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes β€” full credit and traffic to the original publisher.