
On this page
- How Shopify dynamic pricing actually works
- Rule-based vs AI dynamic pricing
- Shopify price lists, Markets and B2B catalogs: what you get natively
- What Shopify Plus adds for pricing logic
- Where an AI dynamic pricing engine plugs in
- Margin guardrails we put on every pricing engine
- Is dynamic pricing legal? The rules as of November 2026
- Customer trust risks the law does not cover
- AI pricing tools for Shopify: buy, build or skip
- Frequently asked questions
- What we’d do first
Shopify dynamic pricing is possible on every plan, but Shopify does not ship a repricing engine. Prices move when something writes them: an app or script calling the Admin API, a Discount Function at checkout, or, on Plus, a Cart Transform function that overrides the line price. The engine that decides the number, rules or a model, sits outside Shopify.
That split matters more than which app you pick. Most pricing failures we get called in to fix are not bad algorithms. They are decent algorithms with no floor, writing unreviewed prices into a store that also runs three discount apps. The law moved in 2026 too: New York requires a disclosure when personal data sets a price, Connecticut has passed a ban starting in 2027, and the FTC proposed an enforcement policy in August.
Key takeaways
- Shopify dynamic pricing has no built-in engine. You combine stored prices (Markets and B2B catalogs), checkout-time logic (Discount Functions) and, on Plus, Cart Transform line updates.
- Start with rules. AI pricing earns its cost on large catalogs with real demand data, not on 40 SKUs.
- Every engine needs a margin floor, a maximum move per day and a kill switch before it touches a live price.
- Pricing by demand, stock or competitor moves is broadly legal. Pricing by who the shopper is now triggers disclosure rules in New York and bans in parts of the US from 2027.
- In the EU, frequent price moves lower your 30-day “prior price”, which limits the discount you can advertise later.
How Shopify dynamic pricing actually works
On Shopify, a dynamic price is either a stored price that something keeps rewriting, or a price adjusted at the moment of purchase. Most stores that do this well use both.
Stored prices are the variant price, compare-at price, and any per-market or per-B2B-catalog price. Every channel that reads your catalog, feeds included, sees them. A pricing app or custom service updates them through the Admin GraphQL API (in practice, the productVariantsBulkUpdate mutation).
Checkout-time adjustments leave the stored price alone and change what the buyer pays in the cart, either as a visible discount (Discount Functions) or by replacing the line price (Cart Transform). They suit “this cart, right now” logic and are invisible to Google and marketplaces.
Our rule of thumb: if the price should be public and the same for everyone, write the stored price. If it depends on the cart, adjust at checkout.

Rule-based vs AI dynamic pricing
Rule-based pricing changes a price when a condition you wrote becomes true. AI dynamic pricing estimates how demand responds to price and picks the number it expects to earn most within your limits. Rules are cheap and predictable. Models find money rules miss, but only with enough sales history.
| Rule-based | AI / model-based | |
|---|---|---|
| Logic | “If stock under 20 and sell-through over 60%, raise 5%” | Estimated price elasticity per product or cluster, then optimise for margin or revenue |
| Data needed | Inventory, cost, maybe one competitor feed | Months of orders at varied price points, traffic, stock, seasonality, competitor prices |
| Best fit | Clearance, seasonal markdowns, MAP-bound brands, B2B tiers | Large catalogs (thousands of SKUs), commodity or price-compared products, fast-moving stock |
| Main risk | Two rules fighting each other | Opaque decisions, learning from its own price changes, drifting below margin |
Our position: a store with a few hundred SKUs that customers do not price-compare should not buy an AI pricing engine. A markdown calendar and a stock-based rule capture most of the value. AI pays when you have thousands of SKUs, competitors who move daily and enough volume per product to measure a change. Even then, rules are the fence around the model.
From our team: before any model, run one honest test. Hold price on 10 products, move 10 similar ones for four weeks, and compare margin per session, not conversion rate. If you cannot run that cleanly, you are not ready for a model.
Shopify price lists, Markets and B2B catalogs: what you get natively
Shopify gives you four native tools for price variation. None changes prices on a schedule; they give an engine somewhere to write to.
| Mechanism | What it changes | Plan | Good for |
|---|---|---|---|
| Markets catalogs | Stored price per market: a percentage adjustment plus fixed prices per product | All plans | Country or region pricing |
| B2B catalogs | Stored price per B2B market or company, with quantity rules and price breaks | All plans, 3 catalog assignments; company catalogs on Plus | Wholesale tiers, contract pricing |
| Discount Functions | A discount on products, order or shipping, calculated at checkout | Public apps on any plan; custom apps on Plus | Tiered, bundle, customer-tag and cart-based offers |
| Cart Transform (update) | The line price itself, in the cart and checkout | Plus (and dev stores) | Contract prices, configurable products, price tests |
Markets catalogs
Under Markets > Catalogs you set a percentage adjustment per market and override single products with a fixed price, which always wins. Bulk fixed prices go in by CSV through Adjust prices, per Shopify’s price adjustment docs. The catch: the percentage also moves compare-at prices unless you switch off Include compare-at price, which matters if your engine writes compare-at values. Market setup is in our Shopify Markets guide.
B2B catalogs and volume pricing
Quantity rules, price breaks and net terms now work on every plan. Below Plus you get three active catalog assignments across B2B markets, and per Shopify’s B2B plan comparison one catalog in two markets uses two. Company-level catalogs stay Plus only. For most wholesale programs, volume breaks are all the “dynamic” pricing needed (setup is in our Shopify B2B guide).
Discount Functions
The Discount Function API covers product, order and shipping discounts in one schema, and reads customer tags, cart attributes and metafields. That is how most “VIP price” and “buy more, save more” logic is built. A store can activate at most 25 discount functions, and they run concurrently without knowing about each other, so test every combination you allow.
Cart Transform
The Cart Transform API can expand a line, merge lines into a bundle, or update a line’s price, title or image. The update sets a fixed per-unit price in the buyer’s currency and only works on Plus or development stores. Each app gets one cart transform per store, and checkout rejects these operations on subscription lines.
What Shopify Plus adds for pricing logic
Plus matters for pricing in two places: custom Functions and line-price overrides. Stores on any plan can use public App Store apps that contain Functions, but per Shopify’s Functions documentation, only Plus stores can run custom apps that contain them. If your pricing logic is your own and you do not want it in a public app, that is a Plus requirement. The second is Cart Transform’s update, the cleanest way to show a contract or test price as the real line price.
Functions also have hard limits. For carts up to 200 lines a function gets 11 million instructions, 128 kB of input and 40 kB of output. It cannot call your model on every cart: the only outbound call is the Discount API’s fetch target, limited to custom apps on enterprise stores with approved network access. So the model runs elsewhere and the function reads precomputed prices from a metafield.
Watch out: if you were on Plus and ran pricing logic in Shopify Scripts, it stopped working when Scripts were retired on June 30, 2026. Any Ruby line-item script that set tiered or customer prices needs rebuilding as a Discount Function or Cart Transform. Our Scripts to Functions migration guide maps each script type to its replacement.
Where an AI dynamic pricing engine plugs in
An AI pricing engine should never write directly to your store. It proposes, a guardrail layer accepts, clamps or rejects, and only then does a writer push to Shopify. That middle layer is the one people skip. The build we usually recommend:
- Signals. Orders and inventory from webhooks, landed cost from your ERP or a metafield, competitor prices, traffic. Leave customer-level personal data out unless legal has signed off.
- Engine. Rules, a model, or a model inside rules. It outputs a proposed price per variant with a reason code.
- Guardrails. Floor, ceiling, maximum move per day, change frequency, MAP and EU prior-price checks. Anything rejected goes to a review queue.
- Writer. Batched Admin API updates on a schedule, plus a metafield of precomputed prices for any Function that needs them at checkout.
- Log. Every proposed and written price, with inputs, for disputes and regulators.
This is the kind of system our shopify ai solutions team builds, and the model is usually the smaller part of the work.
Margin guardrails we put on every pricing engine
A guardrail is a hard limit the engine cannot override. Without them, Shopify dynamic pricing can mark your catalog down to cost overnight on one bad competitor feed. Ours:
- Margin floor. (landed cost + shipping) ÷ (1 − minimum margin − payment fee rate). Recompute it whenever cost changes.
- Ceiling. Usually MSRP. Above it, a shortage price looks like gouging.
- Maximum move and frequency. We start at 5% and one change per variant per day.
- Freeze windows. No changes during a campaign or email send that advertises a price.
- Kill switch. One setting that stops all writes and restores the last approved prices.
A worked example (hypothetical)
Say a product sells at $79. Landed cost is $42, shipping costs you $6 per order, card fees run about 3% of price, and you want at least a 25% contribution margin. The floor is $48 ÷ (1 − 0.25 − 0.03) = $66.67. A competitor drops to $65 and the engine proposes $64.
Without guardrails, $64 goes live below your margin target. With a 5% daily cap and .99 rounding, the price steps $79 → $75.99 → $72.99 → $69.99, then stops at $66.99 on day four because the next step would cross the floor. If the competitor was only clearing stock, you gave up a little margin for four days, not all of it for a month.
function guardPrice({ proposed, current, landedCost, shipping,
minMargin = 0.25, feeRate = 0.03, maxMove = 0.05 }) {
const floor = (landedCost + shipping) / (1 - minMargin - feeRate);
const low = Math.max(current * (1 - maxMove), floor);
const high = current * (1 + maxMove);
const clamped = Math.min(Math.max(proposed, low), high);
return Math.floor(clamped) + 0.99; // simplified: use integer cents in production
}

Want guardrails around the pricing you already run?
Send us your catalog size and the apps that touch price today, and we will map where rules or a model would fit.
Talk to our pricing team
Is dynamic pricing legal? The rules as of November 2026
Changing prices based on demand, stock, time or competitors is legal in the US, UK and EU. The risk sits in prices personalised with a shopper’s data, and in discount claims that do not match your price history. Both got stricter in 2025 and 2026. This is a summary, not legal advice.
US surveillance pricing laws
Since November 10, 2025, a New York business using an algorithm fed by personal data to set a price must show the line “THIS PRICE WAS SET BY AN ALGORITHM USING YOUR PERSONAL DATA” with that price. Greenberg Traurig’s September 2026 survey of state pricing laws tracks that law and what followed:
- Connecticut (P.A. 26-130) generally bans surveillance pricing by retail sellers from July 1, 2027, and requires other online businesses using it to disclose it.
- Maryland bars large food retailers and food delivery platforms from using dynamic pricing or personal data to raise food prices for an individual, effective October 1, 2026.
- New Jersey makes surveillance pricing of groceries unlawful from August 1, 2027.
- California AB 325, in force since January 1, 2026, bars using a shared pricing algorithm as part of an agreement that restrains trade.
- New York’s One Fair Price Act passed in June 2026 and would replace disclosure with a ban. It was not yet law in early September, so check its status.
The FTC’s personalized pricing policy
On August 19, 2026 the FTC proposed an enforcement policy statement on personalized pricing. Per Skadden’s summary, implying a personalised price is the same for everyone may be deceptive, as is presenting a higher personalised price as a “loyalty” deal. Adequate disclosure says the price is personalised, on what basis, and from which kinds of data. It is a proposal, not a rule, but it shows how the agency reads Section 5.
EU: Omnibus prior-price rule and personalised prices
Under Article 6a of the Price Indication Directive, added by the Omnibus Directive, any price reduction claim must show the lowest price you charged in the 30 days before the discount. The rule covers strike-through prices and general “Sale” claims, as Osborne Clarke’s briefing sets out, and the same briefing notes EU law also requires telling consumers when a price was personalised. EU law does not ban dynamic pricing, and the Digital Fairness Act consultation is looking at “starting from” prices that dynamic pricing makes unrealistic.
The catch for an engine: every price cut lowers your 30-day reference. Run it through October and your EU Black Friday “was” price is the lowest price it hit. Store a rolling 30-day minimum per variant per market and have the theme read that, not compare-at.

Customer trust risks the law does not cover
The fastest way to lose trust is to let a customer see two prices for the same thing, which usually comes from bad plumbing, not bad intent:
- Price flicker. A page cached at one price and a cart calculated at another. Shoppers screenshot it.
- Post-purchase drops. Someone buys at $79 and sees $69.99 two days later. Set your price-match policy before launch, not after the first complaint.
- Feed lag. The shopping feed or marketplace listing updates hours after your price does, so ads show a price you no longer charge.
- Logged-in vs logged-out gaps. Even from an innocent member rule, it looks like personalisation. Label member prices.
Our rule for retail stores is simple: price by product and moment, not by person. Changing a price for everyone because stock is low is easy to explain. Changing it for one shopper because of their browsing history is now a legal question in several states, and a reputational one everywhere.
AI pricing tools for Shopify: buy, build or skip
Buy an app when your problem is common, build when the pricing logic is your edge, and skip it when your catalog is small and rarely price-compared. Most stores asking us about Shopify dynamic pricing are in the third group.
Buy
Pricing apps fall into three groups: competitor monitors with repricing rules (Prisync is one), rule and scheduling apps for bulk changes and markdowns, and enterprise platforms sold on custom quotes. Ask each one: can it enforce a per-variant floor from cost, does it respect Markets catalogs or only the base price, does it log every change, and can one switch stop all writes?
Build
Build when pricing depends on data no app sees: ERP contract terms, production capacity or a demand model you want to own. On Plus the checkout logic can live in a custom Function; elsewhere the engine writes stored prices through the API. Where AI pays back more broadly is in AI in ecommerce: 12 use cases.
Skip
Under a few hundred SKUs with stable costs, an engine adds risk you will not earn back. A quarterly price review and a clearance rule will do more.
Frequently asked questions
Can Shopify change prices automatically?
Yes, but only through an app or integration, because Shopify has no built-in repricing engine. Prices change automatically when software writes them through the Admin API, when a Discount Function applies at checkout, or, on Shopify Plus, when a Cart Transform function overrides a line price. Markets and B2B catalog prices follow your rules but never change on a schedule by themselves.
Is dynamic pricing legal for ecommerce?
Yes, changing prices for everyone based on demand, stock or competitors is legal in the US, UK and EU. Personalised pricing is different: New York requires a disclosure when personal data sets a price, Connecticut bans most retail surveillance pricing from July 2027, and EU shoppers must be told a price is personalised. EU discount claims must also show the lowest price of the previous 30 days.
Do I need Shopify Plus for dynamic pricing?
No, most dynamic pricing works on any Shopify plan. Apps can update stored prices through the Admin API, and public apps with Discount Functions run on every plan. Plus is needed for custom apps containing Functions and for Cart Transform’s update operation, which replaces the line price instead of showing a discount.
What is the difference between dynamic pricing and personalized pricing?
Dynamic pricing changes the price for everyone based on conditions like demand, stock or competitor prices. Personalized pricing sets a price for one shopper from data about them, such as location or purchase history. The second carries most of the legal risk, triggering New York’s disclosure rule, the FTC’s proposed policy and the EU personalisation notice.
How often should a Shopify store change prices?
For most stores, no more than once per product per day, and weekly is often enough. Frequent changes cause cached-page mismatches, feed lag and annoyed repeat buyers, and in the EU they lower the 30-day prior price you must show with any discount. Price-competitive catalogs can move daily if each move is capped at a few percent.
What is the best AI pricing tool for Shopify?
The best AI pricing tool is the one that fits your catalog size and data, and for many stores that means no AI at all. Competitor repricing apps suit large, price-compared catalogs, rule apps suit seasonal and clearance pricing, and custom engines suit stores with ERP contract prices or their own demand data. Insist on cost floors, change logs and a kill switch.
What we’d do first
Do not start Shopify dynamic pricing with the algorithm. Start with the three numbers it needs for every variant: landed cost, minimum margin and the highest price you are willing to show. If those do not exist in Shopify today (in metafields or a synced cost field), that is week one.
Then audit what already touches price: discount apps, automatic discounts, Markets catalogs, any Flow that writes prices. On most stores we review, two of those overlap already.
Run rules with the guardrails above on a small product set for a month, logging everything. If margin per session improves, widen it. Add a model only inside those rules. And keep personal data out of retail pricing unless counsel has signed off in writing.
We have built Shopify systems for brands like EcoFlow, Delsey and Kerbholz. If you want the engine, guardrails and Functions built properly, our Shopify intelligence team can scope it with you.




