Sales Promotion Examples: The Fourteen Mechanics

Updated September 2026 · Written and maintained by the Progression Agency strategy team

Every sales promotion does one of four things: it brings a purchase forward in time, it takes a sale from a competitor, it increases the size of a basket, or it does nothing except give away margin to people who were going to buy anyway. The mechanic determines which. This page sets out fourteen promotion types with what each is genuinely good at, the arithmetic of what a discount costs you, and the specific situations where a promotion makes a business worse.

Why most promotions cost more than they return

A discount offered to everyone is paid to the people who would have bought anyway. The promotions that work are conditional — on a new customer, a larger basket, an off-peak slot — because the condition is what makes the discount buy something.

Key takeaways

Six things to carry away, before the detail.

• A promotion can only pull demand forward, take share, grow the basket, or subsidize people who were buying anyway
• Untargeted promotions do the first and the last, because they reach your most engaged customers first
• Calculate the break-even volume lift before choosing a mechanic: at 50% margin a 20% discount needs roughly 67% more units
• Where the discount equals or exceeds gross margin, no volume increase restores gross profit
• Mechanics that give away cost — gift with purchase, samples, loyalty points — protect margin far better than mechanics that give away price
• Measure incremental units against a holdout or comparable period; total sales during a sale proves nothing
• Watch the three to six weeks after a promotion, because pulled-forward demand produces a trough that gets blamed on something else

Everything below expands those seven points, including the full break-even table, the fourteen mechanics with their margin cost, and the legal requirements that attach to specific promotion types.

What are the main types of sales promotion?

Fourteen mechanics in common use: percentage discount, fixed-amount discount, multibuy, bundle, gift with purchase, free shipping threshold, loyalty points, cashback or rebate, coupon, free trial or sample, financing terms, contest or sweepstake, limited edition scarcity, and referral incentive.

They are not interchangeable. Each has a different effect on margin, a different effect on volume, and a different consequence for what customers expect the next time. Choosing between them by which is easiest to explain is how businesses end up permanently discounted.

Read the third column before the second. In most businesses the decision is dominated by margin cost, and mechanics giving away cost — a gift, points, a sample — are structurally cheaper than mechanics giving away price.

What does a discount actually cost?

Far more than the discount percentage, because you lose margin on every unit you would have sold anyway. The break-even volume lift is the number that matters and it is rarely calculated.

The arithmetic is straightforward. If gross margin is fifty percent and you discount twenty percent, your margin per unit drops from fifty to thirty, which means you need roughly sixty-seven percent more units to make the same gross profit. At a thirty percent margin, the same twenty percent discount requires a two hundred percent volume increase.

‘Impossible’ means the discount exceeds or equals the margin, so no volume increase restores gross profit. This table is worth keeping visible during promotional planning, because the figures are consistently more demanding than intuition suggests and almost no promotion achieves the lift in the right-hand columns.

And the break-even is the optimistic case

It assumes every additional unit is incremental. In practice a meaningful share of sales during a promotion would have happened anyway, which means the real lift required is higher than the table shows, sometimes substantially.

Discounts also move future demand, not just current demand

Customers who stockpile during a multibuy are not in the market next month. A promotion that looks successful in week one and is followed by a trough in weeks three to six has moved revenue rather than created it, and the trough is frequently attributed to something else.

Which promotions actually acquire new customers?

Ones that are targeted so that existing customers cannot use them: first-order codes gated to new accounts, samples, trials, referral incentives, and financing that addresses affordability rather than price.

The gating is the whole point. An untargeted sitewide discount reaches your most engaged customers first, because they are the ones reading your email and following your accounts. The people you wanted to attract are, by definition, the ones paying least attention.

Samples and trials remove risk instead of price

For products where the barrier is uncertainty rather than cost, letting somebody try it solves the actual objection without teaching them a lower price. This is why the mechanic is dominant in categories with genuine trial risk and largely absent where price is the barrier.

Referral incentives use the one channel discounting cannot buy

A recommendation from an existing customer carries credibility no advertisement does, and paying for it after the fact means you pay only for outcomes. The design problem is preventing self-referral, which requires a rule about distinct payment methods or addresses rather than trust.

Which promotions increase basket size?

Threshold-based mechanics: free shipping over a value, multibuy, gift with purchase above a spend, tiered discounts and bundles. All work by making the next unit cheaper than the first.

Free shipping thresholds remain the most reliable of these in ecommerce, and the threshold should be set slightly above the current average order value rather than at a round number. Set too low it gives away shipping on orders that already qualified; set too high nobody reaches it.

Three for two beats thirty percent off

The two can be arithmetically similar and behave very differently. A multibuy requires the customer to buy more to access the saving, so the units move; a percentage discount applies to one unit, so the same margin buys a smaller volume effect.

Bundles let slow stock travel with fast stock

Pairing a slow-moving item with a popular one at a combined price moves inventory that would otherwise be marked down harder later. The risk is anchoring: bundle a premium item too often and the bundle price becomes its real price.

How do you choose the right mechanic?

Start with the outcome you want, then pick the mechanic that produces it, then check the arithmetic supports it. Most promotions are chosen in the opposite order.

The last row is the one businesses get wrong under pressure. Matching a competitor’s discount starts a contest decided by whoever has the deeper margin, and responding with added value rather than lower price refuses the contest without conceding the customer.

When does a promotion make things worse?

When it is untargeted, perpetual, used to hit a number, applied to a product that has not established a price, or used as a substitute for demand generation.

The perpetual sale is the commonest failure

A business that is always running something has not discounted; it has repriced. Customers learn the real price within a few cycles, the promotion loses its effect, and the only remaining options are deeper discounts or a painful return to full price.

Never anchor a new product with a deep discount

The first price a market sees becomes the reference. A launch discount that is too aggressive establishes the discounted figure as the product’s value, and recovering from that is far harder than launching at the intended price with a value-add instead.

Promotions cannot fix demand generation

If not enough people know the product exists, a discount lowers the price for the small number who do. The problem is awareness, and it is addressed by demand generation work rather than by margin.

What are the legal considerations?

Reference pricing, free claims, sweepstakes and automatic renewals are all regulated in the United States, and several are actively enforced.

The main areas to check before running anything: a ‘was’ price must reflect a price the product was genuinely offered at for a meaningful period; ‘free’ offers carry specific disclosure requirements; sweepstakes generally require no purchase necessary and clear official rules, with state-level variation; and negative-option or auto-renewal offers carry their own consent and cancellation obligations. The Federal Trade Commission publishes business guidance covering most of this.

The last row is the quiet one. Repeatedly extending a deadline advertised as final is both a credibility problem and, in some framings, a deceptive practice, and customers notice faster than regulators do.

How should promotional performance be measured?

By incremental units and incremental margin against a holdout or comparable period — never by total sales during the promotional window.

• Incremental units: sales above what a comparable non-promotional period would have produced
• Incremental gross margin, which is frequently negative even when units rise
• New customer count specifically, separated from existing customer orders
• Average order value and margin per order, since AOV can rise while margin falls
• Post-promotion trough depth: how far sales fall in the following three to six weeks
• Repeat purchase rate of customers acquired on discount versus at full price
• Redemption rate for coupons and rebates, which determines the real cost
• Whether full-price sales recovered, and how long it took

The sixth measure is the most under-used and the most informative. Customers acquired on a deep discount frequently repeat at lower rates than customers acquired at full price, which means the acquisition cost calculation should include their lower lifetime value rather than just the discount given.

How do promotions fit into the wider marketing program?

As a conversion mechanism at the end of a demand process, not as the process itself. Promotions convert existing interest efficiently and create very little on their own.

The practical sequence is demand first, then conversion. Work that builds awareness and consideration — search visibility, email, paid media — produces the audience a promotion converts. A business running promotions into an audience that is not being replenished is discounting to a shrinking group.

Email is where targeting actually becomes possible

Segment-level targeting is what separates a promotion that acquires from one that subsidizes, and email is the channel where that segmentation is straightforward: lapsed customers get one offer, never-purchased subscribers another, and recent buyers nothing at all.

What are good sales promotion examples for ecommerce specifically?

Threshold free shipping, gated first-order codes, bundles that pair slow stock with fast, post-purchase add-on offers, and lapsed-customer reactivation sent only to a lapsed segment.

• A free shipping threshold set $8-$15 above current average order value, tested in both directions
• A first-order code that only validates on accounts with no prior purchase history
• A bundle pairing a best-seller with an item whose sell-through rate is lagging
• A post-purchase one-click add-on, which converts well because payment friction is already gone
• A lapsed-customer offer sent only to people who have not purchased in a defined window
• A referral code that requires distinct payment methods on both sides
• A limited-run edition where the limit is genuine and stated in units
• A loyalty tier that unlocks free shipping permanently, which converts a recurring cost into retention

These sales promotion examples share one property: each is gated, thresholded or targeted so it cannot be claimed by everybody. That is the difference between a promotion and a price cut, and it is what the arithmetic earlier on this page is protecting.

The post-purchase add-on is the most under-used

Offering something immediately after checkout, when the payment step has already been completed, removes the largest friction in the transaction. It costs nothing to test, does not affect the price of the original purchase, and is invisible to anyone who does not buy.

What are good sales promotion examples for B2B and services?

Different mechanics entirely: pilot pricing, bundled scope, extended terms, waived setup fees, and multi-year commitments in exchange for a rate rather than a discount.

The last row is why straight discounting is rarer and more damaging in services than in retail. A discounted first year sets the expectation for the renewal conversation, and raising the price later requires justifying an increase rather than simply charging the rate.

Pilots need a defined conversion criterion

A pilot without an agreed definition of success becomes an indefinitely extended discount. Agree in writing what result would lead to a full engagement, and on what date that assessment happens, before the pilot starts.

Where do sales promotion ideas usually come from, and why is that a problem?

From what the business did last year, from what a competitor is visibly doing, and from calendar moments. All three sources produce mechanics chosen before the objective.

The better source is the number. Looking at margin, at average order value, at the gap between new and returning customers and at which stock is not moving produces sales promotion ideas that address something specific. Starting from the calendar produces a promotion that has to have a reason invented for it.

Calendar promotions are not automatically wrong

Some are genuinely well matched to demand: seasonal categories have real peaks, and being absent during them concedes volume. The failure is running a calendar promotion in a category with no seasonal pattern because the date exists.

How does consumer sales promotion differ from trade promotion?

Consumer sales promotion targets the end buyer; trade promotion targets the retailer or distributor. They use different mechanics and the second is invisible to shoppers.

Businesses selling through retail need both and frequently fund one at the expense of the other. Trade spend that secures placement without consumer promotion produces stock on a shelf nobody is being told about, which shows up as poor sell-through and threatens the listing.

How long should a promotion run?

Long enough to reach the audience once or twice, short enough that waiting is not a viable strategy. For most ecommerce mechanics that is three to ten days.

Short windows create genuine urgency and limit the margin exposure. Long windows reach more people and progressively convert the promotion into the expected price, particularly if the same duration is repeated. The test is whether a regular customer could reasonably plan their purchasing around your promotional calendar; if they could, the calendar is too predictable.

Predictability is the hidden cost of an annual promotion

A sale that runs the same week every year trains the market to wait for that week. The revenue still arrives, at a lower margin, and the weeks before it become structurally weaker each year as more customers learn the pattern.

Frequently asked questions