A Black Friday customer service checklist for a small store comes down to seven jobs done in order: fix your policies and help pages, plan who covers which hours, write the replies you will send 200 times, cap what AI and per-ticket tools can spend, warn customers about shipping cut-offs, triage the queue during the sale, and get ready for the returns that follow. In 2026, Black Friday is November 27 and Cyber Monday is November 30. The checklist below is dated backwards from those days, with a short version if you are starting late.
It is written for stores with one to ten people answering customers, where the owner is often one of them.
What changes in support during peak weeks
More messages is the obvious part. The mix changes too, and it changes in a predictable order:
- Before the sale: "When does it start?", "Will this item be included?", "Can I get the sale price on what I bought last week?"
- During the sale: discount codes that don't apply, items that sold out mid-checkout, payment errors, and requests to change or cancel an order that was placed ten minutes ago.
- The week after: "Where is my order?", tracking that hasn't moved, wrong sizes and early gift returns.
- December: delay complaints as carriers slow down, then returns and exchanges after the holidays.
Two things follow from that. Most questions are known in advance, so you can write the answers now. And the urgent ones (a checkout bug, a broken code) are mixed in with dozens of routine "where is my order" messages, so someone has to sort the queue, not just work it top to bottom.
The Black Friday customer service checklist at a glance
| When | 2026 date | Task |
|---|---|---|
| Six weeks out | By October 16 | Audit policies, help pages and the sale terms |
| Four weeks out | By October 30 | Staffing plan, shift cover, temp access |
| Two weeks out | By November 13 | Saved replies, AI answers for sale terms, spend caps |
| One week out | By November 20 | Shipping banners, delay notice, test run |
| Sale days | November 27 to 30 | Queue triage, hourly check-ins, incident replies |
| After the sale | December 1 to January 15 | Delay notices, returns wave, post-mortem |
Starting with two weeks or less to go? Do these first, in this order: write the sale terms in one place, write the ten saved replies in step 3, set an AI spend cap and handoff rules, publish your shipping cut-off dates, and agree who is on shift for each day from Thursday to Monday. Everything else can wait for next year.
1. Six weeks out: audit content and policies
Every unclear policy turns into a conversation. Fix the words before the traffic arrives.
- Write the sale terms on one page: start and end time with time zone, what is excluded, whether codes stack, and whether you honor the sale price on earlier orders. Your team and your AI agent will both answer from this page.
- Re-read the returns policy. If you extend the return window for holiday orders, say exactly which orders and until when.
- Check your shipping page against the carriers' holiday dates. A shipping time you advertise is a promise, and in the US the FTC expects you to have a reasonable basis for it.
- Look at last November's conversations, or last month's if you don't have a November. List the top 15 questions and check that each has a public answer.
- Update the help center articles customers reach most: tracking, returns, payment methods, order changes. The help center gap checklist is a quick way to find what is missing.
2. Four weeks out: staffing and shift cover
Peak support is a scheduling problem more than a hiring one. Work out the hours first, then the people.

- Estimate volume per day. Take a normal day's conversations and multiply by what last year's sale did to it. With no history, plan for a few times normal and adjust after Friday morning.
- Turn that into hourly cover. Conversations per hour divided by what one person closes per hour gives the number of people you need on shift. The agent capacity planner does the arithmetic.
- Cover the evenings and the weekend. If your sale opens on Thanksgiving evening or runs through Sunday night, someone has to answer during hours you normally don't work. The after-hours coverage calculator shows what each gap costs you in waiting time.
- Decide who handles what. One person on refunds and order problems, one on pre-sale and chat, and the owner on escalations is a workable split for a team of three.
- Set up temporary help now. Create their logins, give them the sale terms and the saved replies, and have them shadow a real shift before the sale. Check what extra logins cost: on a per-seat help desk each temp is a full seat for at least a month.
Write the rota down with names and hours, and include breaks. On Black Friday the person on chat at 9 pm should not also have started at 7 am.
3. Two weeks out: saved replies and AI answers for sale terms
Write the replies you will send most, once, while you are calm. Ten cover most of a small store's peak:
- Sale start and end times, with time zone.
- Discount code not working (with the three usual causes: excluded item, minimum spend, expired).
- Price adjustment on an earlier order, yes or no.
- Item sold out, with the restock or alternative.
- Change or cancel an order, and the window for it.
- Payment declined or charged twice.
- Order confirmed, tracking not yet available.
- Tracking hasn't moved in three days.
- Holiday delivery cut-off dates.
- How to start a return or exchange.
Keep each one short, put the answer in the first line, and leave a gap for the customer's name and order number so nobody sends a template that is obviously a template. If you need a starting point, the canned responses examples include order and shipping replies you can adapt.
If an AI agent answers on your site, give it the same sale terms page and test it with real questions: "Does the code work on gift cards?", "I bought this on Monday, can I get the Black Friday price?" Fix the source page when it gets one wrong, and retest. On Sellio the AI agent answers on website chat only; email and other channels go to the shared inbox for a person, which is worth knowing when you plan who covers what.
4. Setting AI spend caps and handoff rules
Peak week is when an AI agent saves the most time and when an uncapped bill grows fastest. Set both limits before the sale:
- A spend cap: a daily or monthly message limit, or a maximum monthly budget, so a traffic spike can't turn into a surprise invoice. Sellio lets you cap an AI agent's usage, and AI past the included credits costs extra, so set the cap with the sale volume in mind.
- Handoff rules: the AI passes the conversation to a person for refunds over a set amount, damaged or missing items, double charges, chargeback threats, and anyone who asks for a human twice.
- A fallback message for when the cap is hit or nobody is online, that says when a person will reply.
Check what happens at the cap. A good setup hands the conversation to the inbox and tells the customer a person will follow up; a bad one stops answering mid-conversation.
5. Shipping cut-off banners and delay messaging
In December, many "where is my order" messages really ask "will it arrive in time?" Answer that before they ask.

USPS published its 2026 holiday shipping dates: for delivery by December 25 within the contiguous US, send by December 17 for Ground Advantage and First-Class Mail, December 18 for Priority Mail, and December 19 for Priority Mail Express. Build your own order cut-off from your carrier's date minus your processing time.
Put that date in three places: a site banner, the order confirmation email and a saved reply. For example:
Order by December 15 for delivery before Christmas within the US. Orders after that ship as normal but may arrive after the 25th. Gift cards arrive by email instantly.
If orders will ship late, tell customers before they ask. In the US, the FTC's Mail, Internet, or Telephone Order Merchandise Rule requires a delay notice when you can't ship on time, with a revised ship date (or the reason you can't give one) and the option to cancel for a prompt refund. Write that notice now so it is ready if a supplier or warehouse slips.
6. During the sale: queue triage and priorities
Working the queue oldest-first sounds fair, but it leaves a checkout bug sitting behind forty tracking questions. Sort by impact instead:
| Priority | Type | Target |
|---|---|---|
| 1 | Checkout, payment or code failures affecting many people | Minutes; fix the cause, then reply to everyone at once |
| 2 | Order changes and cancellations still inside the edit window | Before the order ships |
| 3 | Pre-sale questions from someone with items in the cart | While they are still on the site |
| 4 | Double charges, wrong orders, damaged items | Same day |
| 5 | "Where is my order" inside the normal delivery window | Saved reply with tracking, within a day |
Three habits keep the sale days calm:
- A short check-in every two hours: what is spiking, which saved reply needs an edit, who needs a break.
- One person watching for patterns. Five messages about the same code in an hour usually mean the code itself is broken.
- A ready-made incident note for site or payment problems. The incident message templates give you wording you can post in minutes.
A shared inbox helps here mostly by showing who has picked up which conversation, so two people don't answer the same customer differently.
7. After the sale: returns wave and review
Expect tracking questions through mid-December, delay complaints when carriers slow down, and a returns wave in the first two weeks of January.

- Keep one or two extra people on through the second week of December if your volume stays high.
- Send proactive delay notices as soon as you know, using the template from step 5.
- Prepare return and exchange replies for gift recipients, who often have no order number.
- Keep refunds fast. The same FTC rule sets deadlines for refunding cancelled orders, so don't let them queue behind routine questions.
8. Cost surprise check for per-ticket and per-resolution tools
Many support tools bill by usage, and peak week is when usage peaks. Do the math with your expected volume before the sale, not after the invoice. Using list prices checked on September 6, 2026:
| Pricing | Example rate | Hypothetical peak | Extra cost |
|---|---|---|---|
| Per extra ticket (Gorgias) | About $0.36 to $0.40 each | 3,000 tickets over plan | About $1,080 to $1,200 |
| Per AI outcome (Intercom Fin) | From $0.99 | 1,500 AI resolutions | $1,485 |
| Per AI session (Freshdesk, after the trial) | $49 per 100 sessions | 1,500 sessions | $735 |
| Per seat (Zendesk Suite Team) | $55 per agent a month | 3 temps for two months | $330 |
Ask your vendor three questions: is there an included allowance, what does each unit cost after it, and can you set a hard cap? Per-workspace pricing, which is how Sellio charges, removes the per-seat line for temporary staff as long as you stay within the plan's seats. The AI resolution cost calculator runs the numbers for your own volume, and our Freshdesk vs Help Scout comparison shows how two common tools price a small team.
9. Post-mortem template
Hold a 30-minute review in the second week of December, while it is fresh, and add a short follow-up in January for returns. Answer these and keep the notes for next year:
- Volume: conversations per day from November 20 to December 7, by channel.
- Speed: first response time and resolution time on the peak days compared with a normal week.
- Reasons: the top ten contact reasons, and which could have been answered on the site.
- Saved replies: which were used most, and which were edited every time (rewrite those).
- AI: how many conversations it resolved, how many it handed off, and the questions it got wrong.
- Satisfaction: CSAT on peak days compared with a normal week.
- Cost: what support tools cost in November and December against the plan.
- Staffing: which shifts were overloaded and which were quiet.
- One change for next year, with an owner and a date.
The support KPI scorecard and the CSAT calculator help turn those numbers into something you can compare year to year.
Common mistakes
- Writing saved replies on the day. They come out long and inconsistent, and three people write three versions.
- Promising delivery dates the carrier hasn't promised you.
- Letting the AI agent answer refund disputes. Hand those to a person.
- No owner for the queue. Everyone answers, nobody sorts.
- Forgetting the weekend. Saturday and Sunday of the sale are full shopping days.
- Turning on a usage-billed feature for the sale without checking the price per unit.
FAQ
When should a small store start preparing customer service for Black Friday?
Six weeks ahead is comfortable, which in 2026 means mid-October. Two weeks is enough for the essentials: sale terms, saved replies, AI limits, shipping dates and a rota.
How many extra people do I need for Black Friday support?
Divide expected conversations per hour by the number one person closes per hour. Use last year's sale if you have it; otherwise plan for a few times normal volume on Friday and Monday and keep someone on call.
Should I use an AI chatbot for Black Friday?
It can take the routine questions (sale times, code rules, tracking) if it answers from an up-to-date sale terms page. Set a spend cap, test it first, and route refunds, damaged items and angry customers to a person.
What should a Black Friday shipping delay message say?
The new ship date, or why you can't give one, and the option to cancel for a refund. In the US that is what the FTC's mail order rule requires when you can't ship on time.
How do I handle the returns wave after the holidays?
Publish the extended return window clearly, prepare replies for gift recipients without order numbers, and keep refund times short. Review the top return reasons in January and fix the product pages that caused them.




