Peak Season Ecommerce Checklist: Stock, Support and Shipping

Before peak season, a small online store should fix four things: stock cover for each best-selling SKU, the last order dates for each carrier service, a set of support replies for the questions that will flood in, and an ad budget with a stop rule. AI can build the checklist, the sums and the macros in about a week; you supply the real figures.

The order matters. Stock has the longest lead time, so it goes first. Carrier dates come next because they appear on your site and in your emails. Support macros and ad budgets can be written in a day each once the facts are settled. Doing it backwards, with a lovely email calendar and an empty shelf, is the classic way to waste a good season.

Follow me on Instagram@sagnikteaches

The seven-day build, day by day

This plan assumes one person doing two or three hours a day. It fits a store with up to a few hundred SKUs. If you have more, run it on your top twenty by revenue first, which usually covers most of the sales that matter.

Connect on LinkedInSagnik Bhattacharya

Stock cover by SKU

Stock cover is how many weeks of selling a SKU's current stock supports at the expected peak sales rate. The calculation is simple, and doing it per SKU rather than for the whole shop is the point, because a shop-wide figure hides the three products that will sell out in week two.

Subscribe on YouTube@sagnikteaches

For each SKU you need four numbers: normal weekly sales, a peak multiplier from last year (peak weekly sales divided by a normal week), the number of peak weeks, and the supplier's lead time. Then it is arithmetic. Peak demand equals normal weekly sales times the multiplier times the number of peak weeks. Order quantity equals peak demand plus a safety allowance, minus stock on hand and stock already on order.

Two SKUs checked six weeks out

The figures below are invented for illustration. A small store selling home-fragrance refills and gift sets expects six weeks of peak and runs the sum on two SKUs first.

The last two rows are the alarm bell. The diffuser refill has 2.2 weeks of cover at peak pace against a 3-week lead time: if peak started today it would run out almost a week before a new delivery landed, so the order goes in now, not when the shelf looks low. The gift set has 2.5 weeks of cover against a 5-week lead time: it needs an order now, and if the supplier's minimum is 200 units the store has to decide whether to take 200 instead of 162. The sheet makes the decision visible. It does not make it for you.

Prompt for the model, with the sheet pasted as a table:

Here is a table of SKUs with columns: SKU, normal weekly sales, last year's peak
weekly sales, peak weeks, stock on hand, stock on order, supplier lead time (weeks),
supplier minimum order quantity.

For each SKU calculate: peak multiplier, peak demand, safety allowance (one peak
week), order quantity, and weeks of cover at peak rate. Flag every SKU where weeks
of cover is below lead time. Show the arithmetic for the three largest orders.
Do not invent data; if a column is blank, say so and skip that SKU.

Check the output by hand for the top three SKUs. Models sometimes divide in the wrong direction or round early. If last year was a stock-out year, last year's sales understate real demand, so add a note beside those SKUs and use judgement, perhaps a higher multiplier. If a product is new this year, there is no history, and the honest approach is a small first order and a plan to reorder in-season if the supplier can do it.

Getting last year's sales export into shape

Most peak plans stall on day one because the data lives in three places. Export orders by line item from your store platform for the same period last year, including cancelled and refunded lines, because those tell you what customers wanted but did not keep. Export current stock levels, and add a column for supplier lead time that you fill in by hand from emails or your supplier portal, since platforms rarely hold it.

Ask the model to restructure the export into one row per SKU per week. Illustrative target row: SKU RF-500, week 47, units 96, refunds 4, stock-out days 0. The refund and stock-out columns are the useful ones. If a SKU shows three stock-out days in its best week, last year's sales are a floor, not a ceiling, and you should mark it for a higher multiplier.

I will paste an orders export with columns: order date, SKU, quantity, refunded quantity.
Return a table with one row per SKU per ISO week showing units sold, units refunded and
net units. Then list the 20 SKUs with the highest net units over the peak weeks
[week numbers]. Do not estimate missing weeks; show them as blank.

A third SKU shows how to treat the awkward cases. Say a new product launched in October with six weeks of data and no peak history. Do not apply last year's multiplier from a different product. Take its six-week average, apply the multiplier from the most similar existing product (a gift set behaves like another gift set), mark the figure as a guess, and order the smaller of the supplier minimum and eight weeks of that estimate. Then plan one reorder in-season and ask the supplier whether a rush batch is possible and what it costs.

Carrier cut-off dates

A carrier cut-off is the last day an order can go in order to arrive by a promised date. Your site then needs a shipping promise your warehouse can keep. The information has to come from your carriers' own pages or your account contact, for the exact services you use, and it changes each year. Do not take dates from a general article, and do not let a model produce them from memory.

Build the table yourself and let the AI turn it into wording. Columns: carrier and service, destination zone or delivery area, last order date for arrival by the promised day, your own packing cut-off time, and the date you checked it.

A filled-in row, with the dates left as placeholders because only your carrier can supply them, reads: "Carrier A, tracked 48-hour service, main delivery area, last order [date from the carrier's page], our packing cut-off 1 pm, checked [date] by [name]." One row per service you actually sell at checkout. If you offer a premium next-day option, it gets its own row with its own later cut-off, and the banner should mention it only if the warehouse can really pack that late on the busiest day of the year.

Here is my table of carrier services and last order dates (checked [date]).
Write:
1. A 20-word banner for the top of the site.
2. A longer paragraph for the shipping page, with the table turned into a list.
3. One sentence for the order confirmation email.
Use only the dates in the table. Add a line saying dates may change and to check
the shipping page. Plain English, no exclamation marks.

Illustrative output for the banner: "Order by [date] for standard delivery before [date], by [date] for express. Dates last checked [date]." The model should not add anything about guaranteed delivery. If a carrier does not guarantee dates, your site should not either, and your terms should say what happens if something is late.

Add one internal rule: the day before each cut-off, someone checks the order queue and packs anything that would miss it. That is a calendar entry, not a system.

Support macros built from last year's questions

Peak support is about volume. The same twelve questions arrive a hundred times each, and every minute you save on them is a minute for the unusual ones. Export last year's support emails or chat transcripts for the busiest six weeks, remove names, emails and addresses, and ask the model to group them.

Below are 150 customer messages from last year's peak, with personal details removed.
Group them into themes and count each theme. For the ten biggest themes, write a reply
macro under 90 words, using the placeholders [first name], [order number] and [date].
Do not promise refunds, delivery dates or compensation; use [option] where a decision
is needed from staff. Plain, kind tone.

Typical themes for a small store: where is my order, can I change the delivery address, can I cancel, the parcel shows delivered but I do not have it, the item arrived damaged, wrong item sent, is this in stock again, gift receipts and returns. The model's first batch will be too long and too apologetic. Edit them down by hand and add the facts only you know.

An illustrative macro for "where is my order" after editing: "Hi [first name], thanks for your order [order number]. It left us on [date] with [carrier], and the tracking link is [link]. Parcels in busy weeks can show no movement for a couple of days, so please check again by [date]. If it still has not moved, reply here and I'll chase it." Notice there is no promise of arrival. The same rule applies to every macro.

Keep a human on the sensitive replies

Damaged goods, late gifts and "parcel shows delivered" claims need a person's judgement, because they come with a refund or replacement decision. Give staff a decision table: what they may approve alone, what needs the owner. Your customer service macro and backlog checklist goes deeper on rotas and backlogs, and the chatbot flow design tutorial covers order-status and returns flows if you want a bot to carry the repetitive load.

Return window wording for the season

Many stores extend returns for gifts bought in the run-up to a holiday: items bought from a start date can be returned until a later end date. The wording matters because customers read it as a promise. Keep it short and fixed to dates.

Write a short returns notice for a peak-season extended window.
Facts: items bought between [start date] and [end date] can be returned until
[return-by date]. Customer pays return postage unless the item is faulty. Items must
be unused and in original packaging. Sale items are [included / excluded].
Output: a 40-word summary for the product page, a 120-word version for the returns
page, and one line for the order confirmation email. No legal claims.

Have the final wording checked against your normal terms, and ask a qualified adviser if you sell where consumer rules give customers rights you cannot reduce. Do not delete your standard policy, link to both and say which orders each covers.

Ad budgets with a stop rule

Peak weeks get expensive. Costs per click rise when many stores bid for the same shoppers, so last year's return is a guide to direction, not a price. Build the budget from margin, not from hope.

Say a product sells at $30 with $12 cost of goods, $4 shipping and packing, and $2 in fees, so contribution before ads is $12. If you spend more than $12 to get one sale you lose money on the first order. Break-even cost per sale is $12. Set the pause trigger a little below it, say $10, so that the margin covers the cost of returns and your time. Give the AI last year's weekly ad spend, sales and margin by product, and ask it to compute cost per sale by week, the break-even and a proposed weekly budget with one pause rule per campaign.

Write the stop rule as a sentence a tired person can follow: "If any campaign's cost per sale is above $10 for two days running, pause it and tell [name]." Then actually check on those days. The Etsy ads tutorial shows the same break-even thinking on a marketplace.

Five more checks that save a bad week

Site and checkout

Test checkout on a phone with a real discount code and a real address, from a different network. A broken code during a sale day is the most expensive bug a small store meets.

Back-in-stock and low-stock signals

Switch on a back-in-stock sign-up for any sold-out product. The back-in-stock flow tutorial shows one way to do it and write the email.

Warehouse and packaging

Count packing supplies against peak order volume, not normal volume: boxes, tape, labels and gift wrap. A mailer shortage on day three means a delay on day four.

Staff and holidays

List who is working each day of the peak and who covers sickness. A three-person team with one person away has lost a third of its capacity, and no macro changes that.

Cash and payment holds

Stock orders arrive before sales cash. Make a simple weekly cash table covering supplier payments, ad spend, refunds and expected sales payouts. If a payment processor holds funds during review, you want to see the gap in advance. The payout reconciliation tutorial explains how payouts and fees line up.

Where peak plans quietly fail

• The multiplier is borrowed from a stock-out year. Last year's sales were capped by empty shelves, so demand looks smaller than it was. Mark those SKUs and add a margin.
• Macros promise dates. One macro says "it will arrive by Friday" and the carrier does not agree. Strip every promise and link to tracking.
• Cut-off dates copied from last year. Carriers change dates. Date-stamp every row and re-check before publishing.
• Ad spend keeps running at a loss. A rule written down but never checked is not a rule. Put a daily five-minute check in the calendar.
• The checklist has no owner. Each job needs a name and a day. A job owned by everyone is done by nobody.

A dry run that finds the gaps

On day seven, place three real test orders from a phone, one with a discount code, one for a gift set to a different delivery address, and one that you plan to return. Follow each order through to the end, and write down every point at which a person had to step in. A typical finding: the confirmation email shows last year's shipping dates, or the returns label link goes to a page that does not mention the extended window. Fix those before the first customer finds them.

Next, read ten of your own support macros as a customer who is anxious about a gift arriving on time. Wherever the reply makes you more worried, rewrite it. Ask the model to mark phrases that sound defensive or vague, then replace each with a fact: a date you checked, a link to tracking, a named next step.

A short before and after for the shipping banner shows the tone. Before: "Order now to get your gifts in time!" After: "Order by [date] for standard delivery by [date]. Dates last checked [date]; see the shipping page for other options." The second is longer and calmer, and it can be proved. If dates move, you change one line.

The final sign-off list

On day seven, walk through this list and tick each item only when you have seen the evidence.

1. Every top SKU has a stock cover figure, an order placed or a reason for none, and a lead time compared with the first peak date.
2. Reorder emails to suppliers sent, with confirmation of dates in writing.
3. Carrier dates table filled in, date-stamped, and banner wording live.
4. Support macros loaded in the helpdesk, with no promises of delivery dates or refunds.
5. Decision table for refunds and damaged goods pinned where staff can see it.
6. Peak-season returns wording live, with start, end and return-by dates.
7. Ad budgets set, with a pause rule and a named checker.
8. Checkout, discount codes and confirmation emails tested on a phone.
9. Packing supplies counted against peak volume.
10. Staff rota and sickness cover written down.
11. A weekly cash table for the peak weeks.

Once peak is over, spend an hour comparing the plan with what happened. Which SKUs sold out, which macros were used most, which cut-off dates mattered. Save that note with the plan, and next year's version takes two days instead of seven. If you want help turning your own stock, carrier and support set-up into a repeatable plan, that is the sort of job my AI implementation consultation covers.

Peak season questions from small store owners

How early should a small store start peak preparation?

Start at least six to eight weeks before your first big sales date, longer if your suppliers have long lead times. Stock orders are the part that cannot be rushed, so do that part first and leave support macros and ad plans for the weeks when the stock is on its way.

Where do I find carrier cut-off dates?

From your carriers' own business pages or account manager, and only for the services you actually use. Dates change from year to year, so copy them into your sheet with the date you checked them and re-check before you publish them as delivery promises on your site.

Should I use AI to forecast my peak demand?

Use it to organise last year's sales and spot patterns, not to invent a forecast from nothing. A simple ratio of last year's peak to a normal week, applied per SKU and checked by you, beats a confident number with no data behind it.

Can I automate support replies during peak season?

Draft macros and an order-status bot for the repetitive questions, then keep a human in the loop for refunds, damaged goods and anything emotional. Read a sample of automated replies daily during the peak so mistakes are caught within hours, not weeks.

How do I word a longer returns window without losing money?

State the start and end dates in plain words, say who pays for return shipping, and say whether the window applies to all items or only gifts. Keep your normal terms for the rest of the year and link both versions, so a customer can see which one applies to their order.

Further reads

• AI for Ecommerce Inventory Management: Reorder Points and Stockouts — Reorder points and stockouts, the maths behind stock cover.
• Etsy Seasonal Trends: A 12-Month Listing Calendar With AI — Plan listings across the year, not only the peak.
• How Small Shops Can Use AI to Forecast Stock and Reorder — A gentler introduction to forecasting stock with AI.
• How to Sync Stock Across Shopify, Amazon and eBay Automatically — Stop overselling when you sell on several channels.
• Chargeback Response Template for Online Stores, Drafted With AI — Be ready for the disputes that follow a busy season.
• AI for Ecommerce Analytics: Shopify Reports, GA4 or ChatGPT First? — Pick the first tool for reading peak-week numbers.
• AI in Warehouse Operations: Trace Pick Errors Back to Their Cause — Turn a month of pick errors into ranked root causes and a short list of fixes to test, using only your own error log and an AI assistant.
• Abandoned Cart Recovery in WooCommerce: Plugin Setup and AI Copy — Install a free cart recovery plugin, set consent and timing, have AI draft three emails, and measure what the sequence really recovers.
• Advance Shipping Notice: Prepare an ASN From Pick Lists With AI — Build a draft ASN from your pick list and carton data, run the mismatch checks against the PO, and send only after the customer's guide is satisfied.
• Candle Troubleshooting Replies: Tunnelling, Sooting, Weak Scent — Cause lists, care wording, replacement rules and reply templates for tunnelling, sooting and weak scent, drafted with AI.
• Forklift Safety Checklist: AI-Built Daily Inspection Forms — Draft a daily forklift inspection form per truck type with AI, plus the defect-reporting flow and the checks a competent person signs off.
• Free Shipping Threshold: Set It From Your Average Order Value — A worked order-export example: find your average order value, price the shipping you give away, and test two thresholds with ChatGPT or Claude.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: general ecommerce operations practice; carrier cut-offs, supplier lead times and returns rules must come from your own carriers, suppliers and sales history. No vendor pricing used.