Promotions are easy to launch because the activity feels familiar. Choose a discount, add a banner, send an email, increase the advertising budget, and watch the orders. The store becomes busy, which makes the promotion feel productive before the economics or customer behavior have been understood.
The problem appears afterward. Revenue increased, but was the promotion profitable? Did it attract customers who would have purchased anyway? Did it move the products the business wanted to move, or did it discount the products that were already selling? Did the extra orders create cancellations, service work, or delivery promises the operation could not support? A larger sales number cannot answer those questions by itself.
I want every promotion to have one clear job. It may be designed to create a first purchase, reactivate previous buyers, support a supplier initiative, improve the economics of a specific product group, or test whether an offer removes a real objection. Once the job is clear, the promotion can be designed, limited, and measured like an operating decision instead of treated like a holiday on the calendar.
A discount is a cost, not a strategy
A discount changes the economics of every qualifying order. That makes it a cost of producing a result, even when it never appears as a line item in the advertising account. If the business gives up margin without defining the result it expects in return, it is spending money without a budget or a target.
That cost can be worthwhile. A well-designed offer can make a first purchase easier, increase the size of an order, create a useful reason to contact a customer, or help a strong product receive the attention it deserves. But the offer should earn the margin it consumes.
Start by writing the job in one sentence: ‘This promotion exists to produce this behavior from this customer group during this period.’ If the sentence only says the goal is more sales, it is not specific enough. Nearly any discount can create some sales. The useful question is whether it creates the sales the business wants at a cost the business can support.
This framing also opens options beyond lowering the product price. The right offer may be a useful accessory, a service improvement, a product bundle, clearer delivery terms, or better buying guidance. When the customer’s obstacle is uncertainty rather than price, another discount may be an expensive answer to the wrong question.
Choose one behavior you want to change
Promotions become difficult to evaluate when they try to do everything. One campaign is expected to acquire new customers, reward existing customers, increase average order value, move slow products, and improve the month’s revenue. Those goals can require different audiences, products, messages, and economics.
Choose the primary behavior first. If the goal is a first purchase, the audience and measurement should separate new customers from people who already know the store. If the goal is a larger order, the offer should make a logical additional purchase useful rather than reducing the price of the item the customer already intended to buy. If the goal is reactivation, the business needs a meaningful definition of inactive and a message connected to that customer’s previous interest.
A promotion can create secondary benefits, but it should be judged by the behavior it was built to change. That discipline keeps a strong result in one area from hiding failure in the area that mattered.
- Name the customer group the offer is intended to reach.
- Define the observable action you want that group to take.
- Choose the products or categories that support that action.
- State what should not be discounted or included.
- Give the offer a start, an end, and one owner.
Record the baseline before the banner goes up
A result only has meaning in comparison with something credible. Before launching, record the normal performance for the same products, audience, and time period as well as the available data allows. At minimum, I want to understand order volume, conversion, average order value, contribution profit, cancellations, returns, and the amount of support the orders usually require.
The comparison will never be perfect. Demand changes by season, traffic sources shift, suppliers alter availability, and competitors make their own decisions. The answer is not to pretend those conditions do not exist. Mark them. A supplier price change, an unusual advertising increase, a major stock problem, or a tracking failure belongs next to the result so it is not later mistaken for the effect of the promotion.
Also separate the promoted products from the rest of the store. Blended performance can make an offer look stronger because unrelated products had a good week, or weaker because another category had a problem. The promotion needs its own operating view.
Set the economic limits before excitement arrives
Promotions create momentum. Orders arrive faster, dashboards move, and the team wants to extend the deadline or add more products. That is the worst time to decide how much margin the business can give up or how much demand the operation can handle.
Calculate the expected order economics before launch. Include the discounted selling price, product cost, freight responsibility, payment fees, advertising, expected cancellations or returns, and any extra fulfillment or support work the offer creates. The purpose is an operating estimate, not a prediction with false precision. You need enough clarity to set a boundary.
Then decide what would cause the promotion to pause, change, or end. A contribution threshold may be one limit. Supplier availability, delivery capacity, support volume, or an unexpected mix of low-quality orders may be another. Revenue does not get permission to overrule every other part of the business.
Make the rules visible to the people monitoring the campaign. If only the person who designed the promotion knows the acceptable economics, the rest of the team can celebrate a result that is quietly creating a problem.
- The minimum acceptable contribution from a qualifying order
- The total promotional cost or order volume the business will support
- The products, locations, or customer groups excluded from the offer
- The operating signals that require a pause
- The person authorized to make that decision
Keep the offer simple enough to understand
A complicated promotion creates two kinds of friction. Customers struggle to understand whether they qualify, and the business struggles to understand which part of the offer influenced the purchase. Layers of codes, thresholds, bundles, exclusions, deadlines, and channel-specific messages can create activity without creating useful knowledge.
I prefer the smallest offer that can test the idea. Use a clear audience, a clear benefit, and a clear condition. Keep the message consistent from the advertisement or email through the product page, cart, and checkout. If the customer has to read a paragraph of fine print to understand the value, the promotion is already asking the offer to overcome its own confusion.
Test the complete path before launch. Confirm qualifying and nonqualifying products, mobile and desktop behavior, code entry, price display, shipping calculations, confirmation messages, and the way cancellations or refunds are handled. A promotion that works in the marketing platform but fails in the cart is not a marketing problem anymore. It is a customer-service problem.
Simplicity also improves the review. When one meaningful variable changes, the result can teach you something. When the price, message, audience, product mix, landing page, and advertising strategy all change at once, a good outcome is hard to repeat and a weak outcome is hard to diagnose.
Plan the customer experience after the order
The promotion may end at midnight, but the orders continue through payment review, supplier processing, delivery, support, and possible returns. If the campaign generates demand the rest of the business cannot serve, the short-term lift can create a longer period of preventable work and disappointed customers.
Before launch, confirm the participating products, current supplier information, expected delivery process, and the messages customers will receive. Give support the offer terms and a direct way to resolve common questions. Decide how the business will handle an order placed near the deadline, a product that becomes unavailable, or a customer who qualifies but does not see the benefit in checkout.
The offer should not require the team to improvise policy while customers wait. Common exceptions deserve an answer before traffic increases. The business does not need to predict every unusual event, but it should know who owns the problem and what can be promised next.
This planning protects the learning too. If a promotion produces high cancellations because the inventory information was wrong, the result says very little about customer demand. Operating failures contaminate the marketing test.
Read the result by customer, product, and time
When the promotion ends, do not stop at revenue and platform return. Compare the intended customer group with everyone else. Look at promoted products separately from products that received an accidental lift. Review contribution profit, order size, cancellations, returns, customer contacts, and delivery exceptions. If the job was acquisition, follow the new customers long enough to see whether they behaved like customers worth acquiring.
Pay attention to timing. A promotion can pull purchases forward from the following week without creating much additional demand. It can also teach regular customers to wait if similar offers arrive too frequently. Compare the period before, during, and after the campaign rather than declaring success at the final minute of the sale.
The purpose of the review is a decision, not a presentation. Run the offer again, change one part, narrow the audience, replace the benefit, or stop using it. ‘It did well’ is not a decision. ‘It profitably increased first purchases in this customer group without raising cancellations, so we will repeat the same structure once more’ is a decision that can be tested.
Record what was learned while the context is fresh. Future promotions should begin with that evidence instead of rebuilding the same assumptions every time the calendar presents another reason to run a sale.
- Did the intended customer group take the intended action?
- What did that action cost after the full order economics?
- Which products and channels produced the useful result?
- What happened immediately after the promotion ended?
- Which operating problems appeared because of the offer?
- What specific decision should change next time?
Use a one-page promotion brief
A promotion does not need a long planning document. It needs one page that prevents the important decisions from disappearing across messages and dashboards. Write the job, audience, desired behavior, offer, product scope, dates, economics, limits, owner, and measurement plan. Add the result and next decision when the campaign is complete.
This small record changes the quality of the work. Marketing knows what it is trying to produce. Operations can prepare for the expected order flow. Customer service can answer consistently. The person reviewing performance can compare the outcome with the original purpose instead of creating a purpose after seeing the numbers.
Start with the next promotion already on the calendar. Before designing the banner or writing the email, ask what business problem the offer is supposed to solve. If the answer is unclear, the promotion is not ready. If the answer is clear, build the smallest offer that can solve it, set the limits, and preserve the evidence needed to decide whether it deserves to return.
Promotions are useful when they create profitable behavior and better information. They become dangerous when activity is allowed to stand in for both.