How to prepare customer service for a flash sale or product launch

Black Friday gives you months of warning and a week of pressure. A flash sale or product drop gives you a few days of warning and two hours of chaos. Everyone arrives at once, the stock runs out, a code misbehaves, and your inbox fills faster than anyone can read it.

Summer is full of these events: mid-season sales, back-to-school offers, autumn collection launches, drops timed to a campaign or a creator collaboration. They are also the best rehearsal you'll get for the holiday peak. This guide is a compressed version of our Black Friday playbook, built for short, sharp events.

Why short events overwhelm support

Flash sales and launches break customer service in their own way:

• Everything arrives at once. The announcement email lands, and hundreds of people have the same question within minutes.
• Stock creates frustration. "It said in stock, then it disappeared from my cart."
• Codes and timing go wrong. A code that starts an hour late, or a banner still showing last week's offer.
• Social amplifies problems. One annoyed comment under the launch post is visible to everyone.
• Approvals slow everything down. Agents wait for the founder to decide whether to honour a price.

A short event needs short, decisive preparation.

Step 1: Forecast the spike

Look at your last sale or launch. How many messages arrived in the first two hours, compared with the number of orders or visits? Apply that ratio to this event's expected traffic and add a buffer if the audience is bigger (a creator collaboration, a paid campaign, a waitlist that grew).

Then plan around three moments:

Moment	What happens	How to prepare
Announcement and first two hours	The biggest spike: timing, stock, codes, sizing	Full team online, AI agent live, someone watching for patterns
During the sale	Steady flow of pre-sale and checkout questions	Fast chat answers on product and cart pages
Last hours and sell-out	"Is it really ending?", "Will it restock?"	Clear end time, waitlist or restock answer ready
The days after	Order changes, shipping times, sizes, returns	Order lookup automation, return rules published

Step 2: Write the launch FAQ before you announce

Nearly every launch question is predictable. Write the answers into one short launch FAQ, put it on the sale page, and add it to your help center:

• exact start and end time, with time zone;
• what is included, quantities, and whether it will restock;
• sizing and fit, with a link to the size guide;
• how the code works, what it excludes, and whether it stacks;
• shipping dates for sale orders;
• whether customers can change or cancel orders, and the returns policy for sale items.

Then train your AI agent on the launch FAQ and test it with 20 likely questions, including a few awkward ones ("I was in the queue and it sold out, can you hold one for me?").

Step 3: Decide what agents can do without asking

Waiting for approval is what turns a two-minute fix into an angry customer. Write down, in advance:

• Code problems: agents may apply the discount manually or refund the difference when the code should have worked.
• Out of stock after payment: offer a refund, an alternative, or a priority restock, in that order of preference.
• Wrong size chosen: allow a free size change before shipping.
• Limits: the maximum goodwill amount an agent can offer, and who decides anything above it.

Share the rules with everyone answering messages, including friends and family helping out for the day.

On launch day, a fast and generous answer to a small problem costs less than a slow and correct one.

Step 4: Set up proactive messages

Answer questions before they are asked:

• Before: a waitlist or "notify me" sign-up, plus a reminder email with the exact start time.
• At launch: a short chat message on the sale page ("The sale ends at 8pm CET. Ask us anything about sizing").
• After purchase: a confirmation with a realistic shipping date for sale orders.
• If something goes wrong: a banner on the site and a short post on social, so customers don't have to ask.

Step 5: Staff the first two hours

Most of the pressure lands in the first two hours, so put everyone there:

• One person on chat and social for fast pre-sale answers.
• One person on email and order problems.
• One person watching for patterns and fixing causes: the code, the banner, the stock count. In a small store this is often the founder.

Let the AI agent handle the repeat questions so people can focus on problems and sales.

Step 6: Watch the first 30 minutes closely

Problems show up as patterns. If five customers in ten minutes say the code doesn't work, stop answering one by one and fix the code. Then:

1. update the launch FAQ and the AI agent's content;
2. post a short note on the sale page and in comments;
3. send the affected customers one message explaining what happened and what you did.

Fixing one cause can save hundreds of messages.

Common mistakes

• Announcing before the FAQ is live. The first wave asks everything at once.
• Vague end times. "Ends tonight" in a store with customers in five time zones.
• No restock answer. "Will it come back?" is the most common question after a sell-out.
• Hiding problems. A visible "we know about the code issue" beats fifty private explanations.
• Forgetting the week after. Order changes and returns arrive once the excitement fades.

A simple launch-day setup

1. Seven days before: publish the launch FAQ, set up the waitlist and announcement emails, and agree the agent rules.
2. Three days before: train Tidio's Lyro on the FAQ, test it with likely questions, and add a proactive message to the sale page.
3. Launch day: full team online for the first two hours, one person on patterns, AI agent on all live channels.
4. The week after: keep order lookup and returns answers prominent, then hold a 30-minute review of what customers asked.

For everyday automation beyond launch days, read our guide on how to automate customer support.