Product promotion strategies: the price, the purchase frequency and the margin decide which routes are open to you

Most lists of promotion strategies are a menu with everything on it, which is why they are hard to act on. The routes available to a product are not a matter of taste or budget size. They are decided by three facts you already know: what the product costs, how often the same buyer purchases it, and how much margin a sale carries. A ninety dollar product bought monthly at a thin margin and a nine thousand dollar product bought once at a fat margin do not share a single promotion route, and running the wrong one is how a budget disappears without an argument about why. Work the three facts first, then pick.

The three facts, and what each one rules in

Fact about your product	If it is low	If it is high
Price per purchase	Promotion has to be cheap per person reached, so it leans on owned channels, retail presence, sampling and word of mouth	Promotion can afford a conversation, so it leans on content, demonstration, referral and direct outreach
Purchase frequency	Each sale has to be won from a new buyer, so reach and discovery carry the work	The same buyer can be sold to repeatedly, so the list and the reminder carry the work
Margin per sale	Every route has to pay back quickly, which rules out anything with a long payback	A long payback is affordable, which opens brand building, events and content programmes

Write the three numbers down before the strategy conversation. Most arguments about which promotion to run are two people assuming different answers to those three without having said so.

The four combinations, and the routes that fit each

Nearly every product falls into one of four shapes, and each shape has a short list of routes that work and a shorter list that reliably wastes money.

• Cheap, bought often, thin margin. Food, drink, consumables, small accessories. The job is being available and being remembered at the moment of choosing. Routes that fit: retail and shelf presence, sampling, multibuy mechanics, local reach, a loyalty mechanic that costs you only on a repeat visit, and content that is entertaining rather than explanatory. The route that wastes money here is a long sales process, because the margin on one sale will never pay for a conversation.
• Cheap, bought once or rarely, thin margin. Single purchase low value items, one off services. The hardest shape, because each sale must be won fresh and none of them pays for much. Routes that fit: search, where the buyer declares the intent themselves, marketplaces and listings where the demand already gathers, and referral from an adjacent business that meets the buyer first. The move that fixes this shape is usually a product change rather than a promotion change: add a repeat element, a larger pack or an attached service, and the shape becomes one of the other three.
• Expensive, bought once or rarely, fat margin. Capital equipment, major home work, professional engagements. The buyer researches, compares and needs to trust somebody. Routes that fit: content that answers the comparison properly, demonstration and proof, referral from people who already bought, direct outreach to a named list, and search on the specific phrases of somebody in the middle of deciding. Mass reach is the waste here, because almost everybody who sees it has no need this year.
• Expensive, bought repeatedly, fat margin. Subscriptions, professional services on retainer, business supplies. The best shape to be in, and the promotion priority is retention before acquisition. Routes that fit: the list you already have, onboarding that gets a new buyer to a result quickly, a referral mechanic aimed at existing customers, and account level outreach. The waste here is spending on new buyers while existing ones leave, which is a leak no acquisition budget fills.

New product against established product, which changes the order

The shape tells you which routes are open. Whether the product is new tells you the order to open them in.

• A new product needs the informing job done first. Nobody can prefer something they have not heard of, so the early spend goes at reach among a narrow, well chosen group rather than a wide one. Pick the smallest audience that could plausibly buy, and be present repeatedly in front of it rather than once in front of everybody.
• A new product needs proof manufactured deliberately. You have no reviews, no case studies and no referrals, so the first cohort of buyers is worth more as evidence than as revenue. Build the asking for proof into the delivery, in the first week, while the experience is fresh.
• An established product needs reminding more than informing. The market knows it exists. Spend goes at being present at the moment of need and at the buyers who already bought once and drifted, which is the cheapest demand in any business.
• An established product in a crowded category needs differentiating. If buyers can name three alternatives, promotion that describes the category rather than your difference helps all four of you equally.

Those four jobs, informing, persuading, reminding and differentiating, are what the promotion function exists to do inside the wider marketing mix, and the function's place beside product, price and place is set out in the role of promotion as a marketing function.

Where incentives belong, and where they quietly cost you

A discount or an offer is a promotion route like any other, and it is the one with a second price attached. A straight price cut teaches the buyer what the product is worth, so the next full price looks expensive. A multibuy pulls forward demand you already had, so the following month is quiet. A gift with purchase carries stock risk. None of that makes them wrong, and all of it means the arithmetic has to include the second cost rather than only the discount. Each mechanic, with a worked example and the cost that follows it, is in sales promotion examples and what each one costs you after the sale.

The rule that holds across all four product shapes: use an incentive when the barrier is hesitation, and use a sample or a demonstration when the barrier is doubt. Discounting a product nobody believes in yet buys you a cheap sale and no second one.

How to test a route before you commit a year to it

• Run one route at a time, long enough to matter. Two routes started together leave you with a result you cannot assign. Long enough means at least one full purchase cycle for your product, which for an annual purchase is not one month.
• Decide the number that counts before you start. Enquiries, first orders, or repeat orders. Write it down, because the temptation afterwards is to report whichever number moved.
• Work out the payback from your own figures. Spend divided by new customers gives your cost to acquire one. Compare that against the margin on a first purchase, and for a repeat purchase product against the margin across the number of purchases an average customer actually makes. A route that fails on the first purchase and passes across the relationship is a route to keep, and that is a decision you can only make with your own repeat rate in hand.
• Keep a record of what you tried and what happened. The reason companies rerun routes that failed is that nobody wrote down that they failed, or why.

Product promotion: the questions that decide the plan

What are the main product promotion strategies?

Grouped by what they do rather than by name: paid reach, search presence where the buyer declares intent, content and demonstration, retail and shelf presence, incentives and offers, referral and word of mouth, direct outreach to a named list, events and sampling, and earned coverage. Which of them is open to you is set by your price, your purchase frequency and your margin, which is why the list is the wrong place to start.

How many promotion routes should a small business run?

Two, done properly, plus the list you own. Three is the point where nothing gets the attention it needs, and the common failure is five routes each receiving a tenth of the effort required to work.

What is the cheapest way to promote a product?

The buyers you already have, by a wide margin, because there is no acquisition cost attached to them. After that, search presence for the phrases somebody types when they are ready, and referral from a business that meets your buyer earlier than you do. All three are cheap because the demand already exists and you are collecting it rather than creating it.

Should a new product be discounted at launch?

Prefer a sample, a trial or a demonstration at launch, and keep the discount for later. At launch the barrier is doubt rather than price, and a launch discount sets the reference price for everybody who arrives afterwards.

How do I promote a product with almost no budget?

Pick the single narrowest group that could buy it, be present where that group already gathers, and build the request for proof into every delivery so the first ten customers produce the evidence the next hundred need. Narrow and repeated beats wide and occasional at every budget, and it is the only thing that works at a small one.

When can a promotion route be judged?

One purchase cycle for your product, plus the time your buyers take to decide. For a weekly grocery item that is a few weeks. For an annual professional purchase it is longer than any quarterly review, which is why those routes need a leading indicator agreed in advance, such as qualified enquiries, rather than being judged on revenue alone.

The route that opens without anybody buying it

Every route above is one you choose, pay for and control. There is a new one that arrives whether or not you plan it: a buyer asks ChatGPT, Gemini, Perplexity or Claude who to use for the thing you sell, and gets back three or four companies with a sentence of reasoning each. Nothing in that answer was promoted. It was assembled from reviews, articles, listings and company pages, which means the promotion that decided it was whatever was published about you over the previous year. Worth reading the answer your own category returns before the next plan is signed off. AI Knows Us asks those buying questions across the main assistants and records who gets named and what the answers were built from.