Gains:
- Ability to produce price positioning scenarios by collecting and summarizing competitor price data with artificial intelligence
- Ability to construct dynamic pricing and repricing logic within the framework of margin, cost and algorithm rules
- Ability to respect platform/legal boundaries in price fixing, misleading discounts and competitor data collection, and leave the final price decision to humans
Price is the most sensitive decision in e-commerce. One penny wrong price will either eat into your profits or drive the customer away to a competitor. In marketplaces where dozens of sellers sell the same product, price competition changes within minutes; Many sellers use systems that automatically update the price (repricing). Artificial intelligence (AI) is a powerful aid in this field: it summarizes competitor price data, generates price positioning scenarios, accelerates the calculation of "what will the margin be at this price?" But let's draw a line from the beginning: price is not a decision the AI will make on its own. It is a business decision where price, cost, margin target, campaign, inventory and competition rules intersect; The final approval always belongs to the human. Moreover, there are legal pitfalls in pricing that AI cannot see.
Let's define the two terms. Margin is the margin of the difference between the selling price and the cost; usually expressed as a percentage. Setting a price without knowing the margin is walking in the dark. Repricing is the automatic updating of the price according to predefined rules (according to competitor price, stock, demand). Dynamic pricing is when the price changes with time, demand and conditions.
Components of price decision
A good price decision evaluates at least five inputs:
- Cost: Product cost, commission, shipping, return rate, advertising expense. This is the floor below which you cannot go below the price.
- Margin target: How much profit do you expect from this product? Loss leader — product that intentionally drives traffic at low margins or high-margin product?
- Competitor price: How much is the same/similar product being sold in the market? Where do you stand in the Buy Box competition?
- Demand and stock: If demand is high/stock is low, price elasticity increases; If stock is large, melt price may be required.
- Algorithm and platform rules: Marketplace price/discount rules, minimum price, misleading discount ban.
The following table summarizes the pricing strategies:
Strategy
logic
when
Risk
Cost + margin
Fixed margin over cost
stable product
Can ignore opponent
Competitor focused
Positioning relative to the opponent
intense competition
Can melt the margin
value oriented
According to perceived value
Differentiated product
Correct location is difficult
Dynamic / repricing
automatic by rule
Many products, fast market
Harm if uncontrolled
loss leader
Low price for traffic
If there is cross selling
Damage caused by wrong product
Competitor tracking and scenario generation with AI
AI can quickly summarize and interpret the competitor price data you have (whether collected yourself or from a permissioned tool). It produces a table like "There are 12 sellers for this product, the median price is here, we are standing here, the one with the lowest price has the longest delivery time." Even more valuable, it generates a scenario: "If I reduce the price by 5%, my margin goes down to here, my Buy Box chances change like this, but I get this close to the loss limit." These scenarios speed up your decision-making; But you give the final price from an eye of cost and strategy.
An important technical point: how you collect competitor data makes a legal difference. Scraping a site's data with an automatic bot, in violation of the terms of use, may cause legal and contractual problems. Most marketplaces provide price data through their own dashboard or permissioned API; Use authorized sources.
Attention: Do not let the AI establish a rule such as "give the lowest price, we will always be the cheapest". The endless downward price spiral (price war) reduces margin to zero. Be sure to include a floor price (going below this price) and margin protection in the repricing rule.
Legal limits of pricing
Price is the area to be most careful with AI because some behaviors are criminal.
Price fixing / cartel: Setting common prices by agreeing with competitors is strictly prohibited in competition law and leads to heavy administrative fines. It is legitimate to observe the competitor's posted price; but coordination with competitors is illegal. Your price decision should be independent and unilateral. If multiple sellers use the same AI/algorithm tool and synchronize their prices, this could even be considered a competitive issue.
Misleading discount: Giving a false discount by showing the "old price" as an amount that was never actually applied is a misleading business practice. In discounted sales in Türkiye, it is mandatory to display the reference price (the lowest price applied during a certain period before the discount) correctly. If the AI produces a suggestion such as "make the old price higher so that it looks more attractive", this is against the law.
Exorbitant price / stockpiling: Excessive price increases and stockpiling during extraordinary periods are subject to administrative sanctions.
Four copyable templates
1) Competitor price summary:
Your role: assistant pricing analyst. Below are competitor prices for a product on the marketplace (anonymous seller codes, price, lead time, score). Data: [table]. Task: calculate median, lowest, highest price; comment on our position ([our price]); Write 3 observations, taking into account delivery and score. Note: I will make the final price decision.
2) Margin and scenario calculation:
Product cost: [X]. Commission: [%]. Cargo: [Y]. Return rate: [%]. Advertising margin: [Z].Current selling price: [F].Task: calculate current margin. Then, table the margin and distance to the loss limit in the scenarios where I reduced the price by 3%, 5% and 8%. Show the calculation steps so I can verify them.
3) Repricing rule draft:
Task: draft a safe repricing rule for the following product group.Constraints: do not go below floor price [T]; margin falling below [%M]; competitor is at most [k]% below the lowest; do not reduce price if stock is low.Output: give the rule set item by item and explain the purpose of each rule.
4) Discount compliance audit:
Check the following discount plan for legal compliance. Plan: old price [E], new price [Y], discount period [S]. Task: evaluate whether the reference (old) price is shown correctly, the risk of misleading discounts and the necessary information. Mark the legal point you are not sure of as [CONSULT LEGAL].
three mini cases
Case 1 — Controlled repricing. An accessories salesman could not manually price 900 products. With YZ, he designed a repricing rule set with base price and margin protection: "no more than 2% down from the competitor's lowest, but the margin goes below 18%, price reduction if the stock is below 5". Prices remained competitive, margins were maintained, no products were sold at a loss.
Case 2 — Price war trap. A vendor set up a rule for AI to "always be 1 TL cheaper than the competitor". When the opposite seller used a similar rule, the price fell below the cost, chasing each other throughout the day; Both sides suffered losses. The spiral stopped when the floor price was set. Lesson: the unlimited "be cheapest" rule destroys margin.
Case 3 — Misleading discount penalty. One store claimed a 60% discount by displaying a high "old price" that it never charged. During the audit, a violation of the reference price rule was detected and administrative sanctions were imposed. The correct way: was to show the real reference price.
Weak prompt / Strong prompt
Weak prompt:
You determine the best price to sell this product cheaper than your competitors.
The model does not know the cost, margin and basis; may "suggest" a price that leads to a loss.
Powerful prompt:
Your role: pricing analyst. Product cost is 148 TL, commission is 12%, shipping is 25 TL, return rate is 6%. Competitor median price is 349 TL, lowest is 319 TL (long delivery). My margin target is at least 20%. My base price is 289 TL. Task: generate 3 price scenarios that comply with the floor and margin constraint; Comment on the margin of each scenario and its impact on Buy Box chances. Show account. I will make the final decision.
Common mistakes
- Asking for a price without giving the cost and margin. A suggestion that leads to harm is made.
- Unlimited "be the cheapest" rule. A price war resets the margin; base price is required.
- Coordination with competitors. Price fixing is a crime; The decision must be unilateral.
- Misleading old price. Fake discounts lead to sanctions; The reference price must be displayed correctly.
- Unauthorized data scraping. Collecting competitor data using illegal bots carries legal risks.
- Leaving the price completely automatic. If repricing is uncontrolled, a single faulty rule will cause great harm.
In summary
Price is the most sensitive decision in e-commerce and AI cannot make it alone. A good price decision evaluates cost, margin, competitor, demand/stock and platform rules together. AI summarizes competitor data, generates margin scenarios, drafts repricing rules; But people give the final price. Be sure to include a floor price and margin protection in repricing. Price fixing (cartel) is a crime; Avoid coordination with the opponent. Misleading discounts and price gouging are sanctioned; Display the reference price honestly.
Application task
Subtract the actual cost, commission, shipping and return rate for a product of yours. Calculate your current margin and three discount scenarios with the "Margin and scenario calculation" template and manually verify each calculation. Then write a rule set with floor price and margin protection with the "Repricing rule draft" template. Finally, check your discount plan for misleading discounts with the "Discount compliance check" template.
checklist
- [ ] I incorporated cost, margin, competitor, stock and platform rules into the price decision.
- [ ] I verified each margin calculation manually; I didn't blindly trust the AI's numbers.
- [ ] I added a floor price and margin protection to my repricing rule.
- [ ] I did not engage in any price coordination with the competitor; My decision is one-sided.
- [ ] I showed the reference price honestly in the sale; I did not give a misleading discount.
- [ ] I collect competitor data only from authorized sources.