What to Do in September if You Sell Online

Amazon published its holiday dates back in July, and one of them lands sooner than most sellers expect. For the October Prime event, the inbound cutoff for FBA shipments using Amazon-optimized splits — the default most sellers are on — is September 16. Minimal splits close a week earlier, and Amazon's warehousing service a week before that.

So the popular plan, "we'll get ready for Q4 in October," is already too late for part of Q4. And it is late for a subtler reason than a missed shipment.

September is the last month when changing something is cheap. In October, peak fulfillment fees start and ad auctions get expensive. In November you cannot fix a listing, a photo or a supplier — you can only sell whatever you set up before.

That makes September a month of small, boring, reversible decisions. Here is the order they are worth doing in.

First, the dates that are already fixed

Everything else in this article is judgment. These are not.

If you use FBA, Amazon published all of its 2026 dates back in July — inbound cutoffs for both October events, and the window when peak fulfillment fees apply. Check them against your own Seller Central before you plan around them: cutoffs vary by shipment type and by fulfillment center, and yours is the version that counts.

Look at the second row: the option most sellers are on closes a week before the one they remember.

On eBay, Shopify, Wix or your own storefront there is no cutoff at all — which is exactly why they get planned last and go wrong. Your deadline there is your supplier's lead time plus the slowest carrier you use, counted backwards from the first day you want to be sellable. If that lands before September 16 too, you have one deadline, not two.

Write both dates down somewhere you will see them. Most of the panic in November is caused by a date nobody put on a calendar in August.

Second, count what you have — and what actually sells

The instinct in September is to restock. The useful move is to first separate the catalog into the part that sells and the part that has been asleep all year.

In most stores the split is brutal. A few dozen products carry the revenue and the rest sit there, quietly taking up ad budget, storage fees and attention. Restocking proportionally — a bit of everything — is how a good Q4 turns into a January full of dead stock.

So: pull last year's Q4 by product, not by month. Which items sold, at what margin after shipping and marketplace fees, and which ones came back. Anything that did not sell in twelve months is not going to be rescued by Black Friday, and it should not get a single dollar of peak-season inventory or a single dollar of ad budget.

The same list tells you what to protect. The handful of products that make your season are the ones whose stock you should be paranoid about, whose photos should be redone if they are weak, and whose competitors you should look at now rather than in December.

Third, fix the listings while a rejection is cheap

A listing that gets rejected in September costs you a day. The same rejection on November 20 costs you the item's whole season, because you will be dealing with orders, not with attributes.

Three things are worth a pass now. Photos: the main image is the one thing a marketplace is strictest about, and the one thing you cannot fix quickly at peak. Attributes: the fields nobody fills — material, size, compatibility, model number — are what both marketplace search and, increasingly, an AI shopping assistant filter on. Stock accuracy across storefronts: if the same item lives on eBay, Shopify and Amazon, the fastest way to ruin a good season is selling the same unit twice.

Fourth, look at your account metrics now, not in December

This is the one most sellers skip, and it is the one with the longest memory.

Marketplace performance metrics are judged over a lookback window — 60 days on Amazon, three or twelve months on eBay depending on your volume. A bad week in December is therefore not a bad week; it is a bad quarter, and the penalties land in January when you are trying to recover. Going into peak season already close to a threshold is the single riskiest thing on this list, and the only one you can fix quietly in September. Where those thresholds actually sit is worth twenty minutes of reading before the volume arrives.

The short version, and the shortest path

September, in order: put the two dates on a calendar, split the catalog into what sells and what sleeps, fix the photos and attributes of the part that sells, and check your account health while a slip is still recoverable.

None of it is difficult. It is just four different screens in four different systems, on a weekend, in a month when nothing feels urgent yet.

That gap — work that is easy but scattered — is what we built SellerClaw for: a team of AI agents that connects to the stores you already sell on, along with your suppliers and ad accounts, and that you talk to in a chat, in plain language. Asked what to restock, it answers from last year's orders and the current stock, not from a general opinion about Q4. Asked to fix the catalog, it works through the drafts rather than handing you a list of things to fix.

You can do the whole checklist by hand, and plenty of good sellers will. But do it in September. Connect one store, ask what sold last Q4 and what has not sold at all this year — the free credits cover that conversation — and you will have the only two lists that matter before the deadlines start passing.