Skip to content

Dynamic pricing and AI

| Article: 3 min read | Marketing

Dynamic pricing means changing prices by rules or a model depending on demand, stock levels, competitor prices, time or other conditions. There are simple rules ("price no higher than the competitor + 3%") and AI models that pick the price themselves. A separate case is personalized pricing, where the price depends on data about a specific person; it is regulated more strictly.

In plain words

Airline tickets and hotels have changed prices for a long time. In retail e-commerce, prices are increasingly changed on principles like "lower it if stock is high", "raise it if a competitor sold out", "the discount depends on category and season". The algorithm only calculates faster than a person; the decision about limits and rules stays with the store.

What the evidence shows

There are no fresh public studies on how much dynamic prices raise profit in online stores. There are only vendor case studies that can't be verified, so there are no figures.

An example of practice from a large market: an Impact Analytics analysis (top 1,000 SKUs in 13 categories) showed that on Amazon Prime Day 2025 only 14% of SKUs got discounts versus 38% in 2024, but the average depth of the discount grew by 8 percentage points; Amazon, in the authors' estimate, used demand and stock forecasting models. This is an analysis by a company that sells such solutions, not a statement by Amazon itself (Impact Analytics, 2025).

Regulation: personalized prices

Since 10 November 2025, the state of New York requires companies that set prices by an algorithm using a consumer's personal data to display next to the price the notice "THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA". Penalties are up to 1,000 US dollars per violation; a federal court dismissed the lawsuit by the National Retail Federation, which considered the law a violation of free speech (the status of the appeal is not given in the source) (Fenwick). The law applies in the US; Ukraine has no regulation directly dedicated to algorithmic pricing. For stores selling in the EU or the US this is a signal: different prices for different people based on personal data is a high-risk zone.

How to start safely

  1. Start with rules, not AI: the lower bound from margin (see pricing), the upper bound from the market.
  2. Change prices based on product circumstances (stock, supplier price, season), not on data about a person.
  3. Keep the same price on the site, in the feed and on marketplaces (Product feed).
  4. Record every change: what was changed and what effect it had on margin.
  5. Set safeguards: a maximum frequency of changes, a ban on going below cost.

Common mistakes

  • Letting the model optimize only conversion: it will lower the price to zero margin.
  • Chasing a competitor's price without knowing their delivery terms and availability.
  • Different prices on the site and in ads: the buyer sees one, the site shows another.
  • Ignoring trust: Baymard (updated September 2025) records that for 12% of shoppers who abandoned a cart for reasons other than "just browsing", the reason was that the total cost was not visible upfront (Baymard).

FAQ

What is dynamic pricing? Automatic price changes depending on demand, stock, competitor prices or time.

Is it legal in Ukraine to change prices dynamically? There is no specific ban on this; the general rules on prices, advertising and consumer protection apply. For your case, consult a lawyer.

Do you need AI for dynamic prices? No, start with simple rules. AI makes sense when you have thousands of items and a lot of data.

Sources

Share

Related

  • What is an e-commerce strategy, who needs it and why projects fail without one

    | Article: 5 min read

    What an e-commerce strategy is and how it differs from a marketing plan, who needs it, why online stores fail without one and how to develop it step by step.

  • Preparing for big sales: Black Friday

    | Article: 3 min read

    Black Friday is a sale that falls on the Friday after American Thanksgiving; in 2026 that is 27 November. Around it forms a "sales week" (Cyber Week) that runs up to Cyber Monday. For a store this is the peak period, when mistakes in stock, delivery and the…

  • Retail media: advertising on marketplaces and in retailers' stores

    | Article: 5 min read

    Retail media is advertising a brand or seller buys on the platform where the purchase happens: a marketplace, a retailer's online store or its app. Ads are shown to people who are already searching for a product or comparing, and the result is visible in that…

  • Selling through social media and messengers in Ukraine (social commerce)

    | Article: 6 min read

    Social commerce is selling goods where people chat and watch content: Instagram, TikTok, Facebook, Telegram, Viber. The customer finds a product in a feed, a video or a chat and can ask questions, choose and agree on payment without leaving the app. In…