01
The only question that matters
A discount is worth it if the additional profit from incremental sales exceeds the margin given away on sales you would have made anyway. That is the entire framework. Everything else is a question of which mechanism concentrates the discount on the incremental buyers and keeps it away from the ones who were already going to purchase.
This is why a site-wide percentage off is usually the weakest form of promotion available. It reaches everyone, including the buyer who had the item in their basket at full price, and the majority of the margin sacrificed buys nothing. A targeted offer to someone who has watched an item for two weeks without buying is aimed precisely at the person the discount might actually move.
02
The mechanisms, ranked by precision
Availability varies by marketplace, but the categories are broadly consistent. They are listed here from most targeted to least.
| Mechanism | Who it reaches | When it earns its cost |
|---|---|---|
| Offer to a watcher or interested buyer | People who have already shown intent and not converted. | Almost always the best value. The buyer self-identified as hesitating on price. |
| Accepting or countering an offer | One buyer, actively negotiating. | When the counter protects your floor. Set the floor before you look at the offer. |
| Multi-buy and combined postage | Buyers already committed to one item. | Frequently excellent: increases order value and spreads fixed dispatch cost. |
| Volume discount on a specific line | Buyers of one product. | When you hold depth you need to clear and the unit economics survive. |
| Category or store-wide sale | Everyone, including full-price buyers. | Occasionally, for clearance or a genuine demand peak. Rarely for routine trading. |
| Permanent low pricing | Everyone, forever. | Only if your cost base genuinely supports it — otherwise it is not a promotion, it is your new price. |
Multi-buy deserves particular attention for resellers because of the postage effect. A second item in the same parcel usually adds little or nothing to your dispatch cost, so a discount funded from that saving can be genuinely free to you and materially attractive to the buyer — the arithmetic is in postage strategy.
03
Handling offers and negotiation
- Set a floor per item before it is listed Calculated from cost, fees, postage, and the minimum margin you will accept. Decided in advance, when you are not looking at a specific buyer.
- Respond promptly Offers decay. A buyer who has waited two days has usually bought elsewhere, and a fast reply is worth more than a clever one.
- Counter rather than decline A declined offer ends the conversation. A counter keeps it alive, and a meaningful proportion of low offers convert somewhere in the middle.
- Never counter below the floor This is the whole point of having one. A sale below your floor is work performed at a loss, and volume does not repair it.
- Bundle instead of cutting where you can Adding an item is often cheaper than reducing the price, and it moves stock rather than only margin.
- Record what was accepted The pattern of accepted offers is your real market price. If everything sells fifteen percent below list, your list price is decoration.
One structural warning: consistently listing high in the expectation of negotiating down damages your search position and click rate, because most buyers never make an offer — they filter by price and move on. You lose the many to negotiate with the few.
04
Seasonal events and sale periods
Major shopping events raise traffic and raise discount expectations at the same time. Whether they are worth participating in depends on your category and your stock position rather than on the event's prominence.
- Participate when you have depth to clear or stock that will be worth less after the season passes. Elevated traffic makes clearance cheaper than it will be in January.
- Participate when your category is genuinely gift-driven, where buyers are shopping on a deadline and comparison shopping less carefully.
- Do not participate reflexively with scarce, sought-after stock. Discounting an item that would have sold at full price during the busiest period of the year is the most expensive mistake in this guide.
- Plan capacity before promotion. A successful event you cannot dispatch on time damages account health, and the metrics penalty outlasts the sales bump — see account health.
- Expect a trough afterwards, and expect elevated returns in the weeks following a gifting peak. Both are normal and both should be in the plan.
The buying side of this — holding the right stock ahead of a peak, funded months in advance — is the harder half, and it is covered in seasonality and demand planning.
05
Paid placement as a form of discount
Promoted or sponsored placement is not usually filed under discounting, but it should be. It reduces your realised margin per sale in exchange for additional visibility, which is structurally the same trade as a price cut with a different distribution.
- Fix the listing first. Paying to send more traffic to a listing that does not convert spends money faster without changing anything. See when listings do not sell.
- Understand what you are charged for and when, because models differ substantially between platforms and between programmes on the same platform.
- Account for cannibalisation. A share of promoted sales would have happened organically, and ignoring that overstates the return — sometimes dramatically.
- Compare the rate against a direct price cut of the same value. Occasionally the price cut wins outright, and almost nobody checks.
- Use it where competition is high and differentiation is low, which is where organic placement is hardest to earn.
- Avoid it on scarce items. If you hold something few others have, buyers will find it without paying for placement.
06
Measuring whether the promotion paid
The near-universal error is judging a promotion by units sold. Units always go up. The question is whether profit did.
- Record a baseline first: units, average realised price, and total margin over a comparable prior period. Without this there is nothing to compare against and every promotion looks successful.
- Measure total margin, not revenue, after fees, postage and the discount itself.
- Estimate the cannibalised share. Even a rough deduction for sales you would have made anyway changes the conclusion more often than not.
- Watch the period afterwards. A promotion that pulls forward demand shows a strong week followed by a weak fortnight, and only the pair together is meaningful.
- Check the returns rate. Discount-driven purchases are less considered and come back more often, and the return costs you the fees and the postage.
- Watch for price anchoring. Frequent discounting teaches your regular buyers to wait, which permanently lowers your realised price. This is the slow damage, and it is invisible in any single promotion's numbers.
That last point is why disciplined sellers discount less than they are tempted to. Every promotion is also a lesson to your audience about what your prices really mean, and a store that is always on sale has simply repriced itself while paying the administrative cost of pretending otherwise.
07
Practice
Exercise
Check whether your last promotion paid
- Take your most recent promotion.
- Compare units and total profit against an equivalent period before it ran.
- If units rose and profit did not, you moved margin to buyers who would have bought anyway.
Check yourself
A sale week moved far more units than usual. Was it a success?
Units are the wrong measure — they always rise, because you paid for them to. The question is whether profit rose, which means comparing margin after the discount and after the fee on the discounted price against what those units would likely have earned anyway. A promotion is justified only if it created sales that would not otherwise have happened.
Where does paid placement belong in this thinking?
In exactly the same arithmetic as a discount. It is margin transferred out of the sale in exchange for visibility, and it should clear the same bar: did it produce sales that would not have occurred? Treating it as a marketing line item rather than as a discount is how sellers end up paying to promote items that were already selling.
Progress is saved in this browser only. No account, nothing sent anywhere.
08
Common questions
Are offers to watchers actually effective?
They are usually the most efficient discount mechanism available, because the recipient has already demonstrated interest and not converted — which frequently means price. The discount lands on someone who might genuinely be moved by it rather than on the whole market. Use a meaningful reduction rather than a token one; a very small cut tends to read as an insult and converts poorly.
Should I price higher to leave room for offers?
Generally no. Most buyers filter and browse by price and never negotiate, so a padded list price loses you visibility and clicks among the majority in order to preserve room with a minority. Price realistically, set a firm floor, and negotiate within it.
How often should I run promotions?
Infrequently enough that they remain unusual. Continuous promotion becomes your price level while adding administrative overhead and teaching buyers to wait for the next one. Tie promotions to a real reason — clearance, seasonal demand, depth you need to move — rather than to a calendar habit.
Is free postage a discount?
Yes, and often an efficient one, because buyers respond to it more strongly than to an equivalent reduction in item price and it improves how your listing reads against competitors showing separate delivery costs. It is only worth it if the cost is genuinely built into the price and the item is not so heavy that it distorts your margin — the calculation is in postage strategy.
Does discounting hurt my search placement?
Selling more usually helps placement, and platforms tend to reward conversion. But repeated price edits can reset accumulated listing signals on some platforms, and heavy discounting distorts your own price history in ways that make future pricing harder to judge. The effect varies by marketplace, so treat any specific claim about ranking mechanics with caution and measure your own results.