A/B Testing Prices on Shopify: Find Your Optimal Price Point

Price is one of the fastest levers you can pull in ecommerce: it shapes perceived value, conversion rate, average order value (AOV), refund rate, and long term customer value. It is also one of the easiest levers to pull incorrectly. A small increase can lift revenue immediately; it can also reduce conversion and harm profitability if your traffic is price sensitive or if your offer is not clearly differentiated.

Running an a/b test prices shopify experiment is the safest way to make pricing changes with evidence rather than instinct. Instead of changing prices store wide and hoping for the best, you split traffic between two (or more) price points, measure the impact on key metrics, and keep the option that improves revenue and profit for your specific audience.

This article covers the practical methodology for shopify price testing: how to choose price variants, set up clean tests, decide which metrics matter, calculate sample size, avoid common pitfalls, and interpret results with confidence. It also includes a repeatable workflow for ongoing price split test ecommerce optimisation, plus notes on testing prices alongside product titles and descriptions using tools such as ConvertLab.

Why price testing works (and why it can fail)

Price testing works because price impacts two competing forces:

• Demand: higher prices usually reduce conversion rate and units sold.
• Value per order: higher prices increase revenue per order and can improve gross profit per order.

The best price is rarely the highest or lowest. It is the price that maximises the outcome you care about, typically revenue per visitor (RPV) or gross profit per visitor (GPV). Two stores can sell the same product and have different optimal prices due to traffic source mix, brand trust, shipping times, competitive positioning, and customer segments.

Price testing fails when the test is not measuring the true business outcome, or when the experiment design introduces noise. Common failure modes include:

• Testing too many things at once: changing price while also changing shipping, bundle composition, or ad targeting makes it hard to attribute results.
• Stopping early: price tests often swing early due to random variation. Ending the test after a day or two leads to false winners.
• Ignoring profitability: a lower price can boost conversion but reduce gross margin enough to reduce profit.
• Inconsistent user experience: showing one price on product pages and another in the cart or at checkout introduces distrust and harms conversion.
• Running the test during unusual periods: heavy promotions, stockouts, or delivery disruptions can distort results.

When to run a price A/B test on Shopify

Pricing tests are most useful when you have reasonable traffic volume and a stable product offer. You do not need enterprise scale, but you do need enough purchases to detect meaningful differences.

Good times to test product prices include:

• After you have baseline conversion data: you know your current conversion rate, AOV, refund rate, and contribution margin.
• When you suspect you are underpriced: high conversion, strong reviews, low returns, and minimal customer objections about price can indicate headroom.
• When you suspect you are overpriced: high traffic but low add to cart and checkout completion, especially from high intent channels, can indicate price friction.
• When costs change: rising fulfilment, manufacturing, or ad costs often require price adjustments, and testing helps reduce risk.
• When you introduce a new product: early cohorts can help you find a sustainable price before scaling spend.

Less suitable times:

• During major promotions: Black Friday week can be tested, but it needs careful design because consumer behaviour changes.
• When inventory is unstable: stockouts or low stock banners can change demand independent of price.
• When you are also changing traffic acquisition: big shifts in ad creative or targeting can shift audience quality and distort results.

What pricing tests should optimise for: choose your primary metric

Price affects multiple metrics simultaneously. Before you launch any experiment, decide what “better” means.

Common primary metrics for shopify price testing:

• Revenue per visitor (RPV): total revenue divided by number of unique visitors. This is often the best default for ecommerce because it captures conversion rate and order value together.
• Gross profit per visitor (GPV): (revenue minus COGS minus variable fulfilment costs) divided by visitors. This is the most financially correct metric for many stores, but it requires accurate cost data.
• Contribution margin per visitor: gross profit minus variable marketing costs, useful if you can estimate channel specific costs and want to optimise for scalable growth.

Secondary metrics to monitor (guardrails):

• Conversion rate: purchases divided by sessions or users. Important, but not sufficient alone.
• AOV: average revenue per order. If price increases, AOV should rise; if it does not, discounting or bundling might be interfering.
• Add to cart rate and checkout start rate: useful to diagnose where price is causing drop off.
• Refund rate and chargebacks: higher prices can increase scrutiny and returns; lower prices can attract lower intent buyers.
• Customer support contacts: track “price too high” objections, delivery questions, and discount requests.
• Repeat purchase rate and LTV: for consumables, a lower entry price could increase first purchase volume and future repeat revenue; you need a longer measurement window.

Rule of thumb: choose one primary metric (RPV or GPV), set guardrails (conversion rate cannot fall below X; refund rate cannot rise above Y), then interpret results accordingly.

Pricing test strategy: how to choose price variants

A price test needs a clear hypothesis and realistic variants. Random price points create random outcomes. A better approach is to anchor your variants in customer psychology, unit economics, and competitive context.

Start with your unit economics

Before any price split test ecommerce experiment, write down:

• COGS per unit (including packaging)
• Fulfilment and shipping costs (your cost, not what the customer pays)
• Payment processing fees
• Average discounts applied (if you commonly run promotions)
• Average refund and replacement costs

This helps you avoid running a test that produces a “winner” on revenue but loses on profit. It also sets boundaries: if your contribution margin becomes negative at the lower price, you should not test it unless you are deliberately acquiring customers for LTV and can afford the payback period.

Use demand sensitivity bands rather than tiny changes

Many merchants test a 1% to 3% change and get an inconclusive result. If you cannot detect a difference, you waste time. A stronger approach is to test price points that are meaningfully different while still plausible to customers.

Examples:

• £29 vs £35 (around a 20% difference)
• £49 vs £59 (around a 20% difference)
• £95 vs £115 (around a 20% difference)

When you find the best band, you can run a follow up test with finer steps if needed.

Consider price endings and perceived value

Price perception is not purely mathematical. Some endings signal “deal” pricing; others signal premium.

• .99 and .95 endings: common for value positioning and promotions.
• Whole numbers: can appear more premium or simpler, depending on category.
• Charm pricing vs rounded pricing: test it if your brand positioning is sensitive to tone.

Be careful: changing from £49.99 to £50 changes both magnitude and ending. If you want to isolate the effect of magnitude, keep endings consistent across variants.

Use competitor context carefully

Competitive pricing data can inform your test range, but do not treat competitors as the target. If you differentiate via quality, warranty, shipping speed, or brand trust, your optimal price may be higher.

Practical method:

• Collect 5 to 10 comparable offers in your market.
• Note not just price; include shipping cost, delivery time, guarantees, bundle contents, and review volume.
• Pick a test range that spans from “comfortably within market” to “premium for your position”.

Decide between single SKU tests and broader pricing programmes

Price testing is easiest on a single high traffic product, but the insights may not generalise across your catalogue. Choose the scope that matches your objective:

• Single product test: best for finding an optimal price for a hero product.
• Category level test: useful when products share similar positioning and demand drivers.
• Store wide pricing strategy: more complex; requires careful segmentation and may have interactions with bundles and discounts.

Test design basics: how to run clean experiments

Price tests are A/B tests, so the core principles apply: random assignment, consistent experiences, and a single primary outcome.

Choose between A/B and multivariate approaches

For pricing, the simplest robust design is A/B:

• Variant A: current price (control)
• Variant B: new price (treatment)

You can test more than two price points, but doing so splits traffic and increases the time to reach a conclusion. If you want three prices, consider:

• A/B/C test: three prices at once; slower but can map demand response quickly.
• Sequential tests: run one A/B test, then test the winner against a new price; faster per test but can be influenced by seasonality if spread out.

If you also want to test titles and descriptions, multivariate testing can uncover interactions, but it increases complexity. Tools such as ConvertLab are designed to test prices alongside copy elements in a structured way, so you can learn whether a higher price needs stronger value messaging to win.

Randomisation and consistent allocation

Two rules matter:

• Random allocation: visitors should be assigned to variants randomly.
• Persistent allocation: the same user should see the same price across sessions, within a reasonable window. Seeing different prices on different visits undermines trust and contaminates behaviour.

If your testing tool uses cookies or local storage to keep assignment consistent, confirm how long the assignment persists and what happens across devices. Perfect cross device persistence is difficult without customer login, but you can still run valid tests at session or browser level as long as the pricing is consistent within the journey and checkout.

Avoid price mismatch across product page, cart, and checkout

One of the biggest practical challenges of a/b test prices shopify is ensuring the price displayed remains consistent from product page through checkout:

• Product page price must match cart line item price
• Cart price must match checkout price
• Discount codes and automatic discounts must behave consistently across variants

If your implementation changes only the displayed price but not the underlying variant price used by Shopify, you risk mismatch and customer complaints. Choose a method that updates the actual price used for checkout, or uses a robust discount mechanism that is correctly applied and tracked.

Pick a test duration that covers buying cycles

Duration should be driven by conversions, not calendar time, but time matters because ecommerce behaviour varies by day of week. As a baseline:

• Run at least one full business cycle: usually 7 days to cover weekday and weekend behaviour.
• Avoid ending the test immediately after a marketing campaign launch or email blast unless both variants received the same exposure mix.
• If the product has a longer consideration period, extend the test so returning visitors are represented.

How to calculate sample size and know when results are “significant”

Pricing experiments can produce small changes in conversion rate but meaningful changes in revenue. You want enough data to distinguish a real effect from noise.

Define minimum detectable effect (MDE)

MDE is the smallest change you care about detecting. Set it based on business impact, not curiosity.

• For conversion rate, an MDE of 5% to 15% relative change is common.
• For RPV, you might care about a 3% to 10% uplift depending on margins and scale.

Smaller MDE requires more sample size. If you set an MDE of 2%, you may need an impractically long test.

Understand statistical power and false positives

Most A/B testing uses:

• Significance level (alpha): typically 0.05, which means a 5% chance of falsely declaring a winner when there is no real effect.
• Power: often 80%, meaning a 20% chance of missing a real effect of the size you care about.

These are conventions, not laws. If a wrong pricing decision is very costly, you can demand stronger evidence by using a lower alpha or higher power, but that increases required sample size.

Practical approach for Shopify merchants

Rather than doing manual calculations every time, you can use an online sample size calculator for conversion rate tests, then sanity check with your traffic:

• Estimate baseline conversion rate for the product or store section you are testing.
• Choose MDE (relative change).
• Compute required visitors per variant.

Example reasoning (illustrative, not exact): if your product converts at 2% and you want to detect a 10% relative change (2.0% to 2.2%), you will need a large number of visitors. If you test a 20% price change, you might see larger conversion movement, which reduces required sample size. This is one reason meaningful price steps often produce clearer results.

Sequential testing and peeking

Checking results daily and stopping as soon as one variant “wins” increases false positives. If your tool reports statistical significance, it may still be vulnerable to peeking unless it uses a sequential method designed for continuous monitoring.

Practical safeguards:

• Predefine a minimum run time (for example, 7 days) and a minimum number of purchases per variant.
• Do not stop early unless the effect is very large and consistent, or you hit a guardrail (for example, conversion collapses).
• If you must monitor frequently, consider sequential analysis approaches; some testing platforms bake this in, while others assume fixed sample sizes.

Implementation options for price testing on Shopify

There are several ways to implement pricing tests, each with trade offs. The best choice depends on your catalogue structure, discounting strategy, theme, and appetite for engineering complexity.

Option 1: Duplicate products (simple but messy)

You can create two products with identical content but different prices, then split traffic between them via routing rules, landing pages, or ads.

Pros:

• Uses Shopify’s native pricing with no checkout complications
• Works with most themes

Cons:

• Splits reviews and social proof unless you use a review app that merges them
• Creates inventory management complexity
• Can confuse analytics and merchandising
• Not ideal for organic traffic, since the two pages may compete in SEO

Option 2: Discount based testing (flexible, watch tracking)

Keep the list price constant and apply an automatic discount or discount code to one variant to simulate a lower price. For higher prices, discount based approaches are less direct; you would need to increase the base price and discount back for the control, which can affect perception if customers see “compare at” or discount messaging.

Pros:

• Checkout accuracy is good because Shopify applies discounts natively
• Fast to implement

Cons:

• Discount messaging can influence behaviour beyond the actual price
• Discount codes can leak between users and contaminate variants
• Attribution can get complicated if customers combine discounts

Option 3: Shopify Markets and customer segmentation approaches (advanced)

If you are testing by country or segment, Shopify Markets and price lists can support different prices. This is not classic A/B testing because segments are not randomly assigned; you risk confounding factors like geography, shipping times, and currency effects.

Pros:

• Native support for different prices across markets
• Useful for localisation strategy

Cons:

• Not randomised; results may not generalise
• Hard to isolate price effect

Option 4: A/B testing apps that modify price consistently (recommended for controlled tests)

A proper shopify price testing tool should help you:

• Randomly assign visitors to variants
• Persist assignments
• Ensure price consistency across product page, cart, and checkout
• Measure outcomes such as conversion rate, revenue per visitor, and AOV

ConvertLab supports experiments on product pages including price, title, and description, helping you test price changes while keeping the rest of the experience stable or deliberately varying copy alongside price when needed. It is useful when you want to understand the combination that maximises commercial outcomes, rather than treating price in isolation.

Step-by-step methodology: run your first price A/B test on Shopify

1) Choose the right product

Select a product that meets as many of these criteria as possible:

• High traffic and steady sales volume
• Stable inventory and fulfilment times
• Consistent traffic sources (no huge swings due to one-off virality)
• Clear value proposition (so you can interpret whether price or messaging is the bottleneck)

Hero products are often best because they generate enough conversions to reach a decision faster, and the business impact is larger.

2) Collect a clean baseline

Before changing anything, record at least two to four weeks of baseline performance for the product:

• Sessions and unique visitors to the product page
• Add to cart rate
• Checkout start rate
• Purchase conversion rate
• AOV and RPV
• Refund rate and customer complaints

This baseline helps you set realistic expectations and detect unusual test period behaviour.

3) Write your hypothesis

A good hypothesis states:

• What you will change (price)
• Why it might work (customer perception, price sensitivity, competitive context)
• What you expect to improve (RPV or GPV)
• What you will monitor as guardrails (conversion rate, refund rate)

Example:

• “Increasing the price of Product X from £49 to £59 will increase revenue per visitor because the product has strong reviews and low return rates; we expect conversion to drop slightly but not enough to offset the higher price.”

4) Choose your variants and traffic split

For a first test, keep it simple:

• Control (A): current price
• Treatment (B): new price

Use a 50:50 split unless you have a strong reason not to. Uneven splits can reduce revenue risk, but they slow learning because the smaller variant takes longer to accumulate data.

5) Decide your primary metric and guardrails

Recommended default:

• Primary: revenue per visitor
• Guardrails: conversion rate, refund rate, and customer support contacts

If you can measure costs reliably, consider gross profit per visitor as the primary metric, especially for low margin products.

6) Audit discounts, bundles, and subscriptions

Shopify stores often have multiple pricing systems interacting:

• Automatic discounts (for example, “10% off when you buy 2”)
• Discount codes from influencers or email flows
• Bundles (native bundles or bundle apps)
• Subscriptions (selling plans)
• Compare at pricing

These can distort your price test. For example, if Variant B has a higher price, a percentage discount code creates a larger absolute discount, which can partially neutralise your intended difference.

Practical options:

• Exclude the product from discounts for the test period.
• Pause non-essential promotions on the product during the test.
• If you cannot pause discounts, track discount usage per variant and interpret results accordingly.

7) Ensure tracking is correct

To interpret a price split test ecommerce properly, you need clean analytics:

• Confirm your conversion tracking is consistent in Shopify Analytics and Google Analytics (if used).
• Make sure your A/B testing tool tracks variant assignment and ties orders to variants.
• Check that taxes, shipping, and currency are handled consistently across variants.

Test the full purchase path for both variants using a test payment method where possible. Verify that:

• The displayed price matches the charged price
• Discounts apply as expected
• Order confirmation data includes the correct line item price

8) Launch and monitor for operational issues

When the test goes live:

• Monitor for errors, theme conflicts, or price mismatch reports.
• Watch customer service channels for confusion or complaints.
• Check inventory levels and fulfilment times; operational disruptions can change demand.

Try not to interpret performance until you have enough data. Early swings are normal, especially for products with lower daily order volume.

9) Analyse results with the right lens

At the end of the planned test window, evaluate:

• Primary metric uplift: RPV or GPV per variant
• Confidence: statistical significance or credible intervals, depending on your tool
• Guardrails: did conversion, refunds, or complaints move beyond acceptable limits?
• Segment effects: did one traffic source or device type react differently?

Segmentation is useful for learning, but be cautious: if you slice data into many segments, some will appear to have large effects by chance. Treat segment results as hypotheses for follow up tests unless you have strong volume.

10) Roll out, then validate post-test

Once you choose a winner, implement it as the default price and monitor performance for at least two weeks. This checks that the test result holds when 100% of traffic sees the new price and when marketing campaigns change.

Also record the learning, not just the winner:

• How sensitive was demand to a 10% to 20% price change?
• Did higher prices require stronger value messaging?
• Did refund rate change?

Documenting these patterns helps you price future products more effectively.

Advanced tactics: improve your pricing experiments

Test price with supporting copy (price and value must match)

Higher prices often need clearer justification. If your Variant B price is higher, consider testing a copy variant that increases perceived value:

• More specific benefit statements
• Clearer materials and quality claims (only if true and supportable)
• Trust signals: guarantees, delivery timelines, reviews, certifications
• Comparison tables or “what’s included” clarity

This is where tools like ConvertLab can be useful: you can run experiments that test price alongside titles and descriptions, so you can learn whether customers reject the higher price itself or the way the value is communicated.

Run tiered pricing tests for bundles and multipacks

If you sell a consumable or a product that benefits from buying multiples, you can test pricing structure, not just a single number:

• Single unit price vs bundle price
• “Buy 2 save 10%” vs “Buy 3 save 15%”
• Bundle anchored to a premium tier (for example, 1-pack, 2-pack, 4-pack)

Primary metric should still be RPV or profit per visitor, because bundles change AOV and conversion simultaneously.

Consider decoy pricing and anchoring (carefully)

Anchoring can shift choices without changing the base product price. Examples include:

• Introducing a higher priced premium option to make the mid-tier feel better value
• Showing compare at prices where appropriate and honest

Be careful with misleading anchors. If customers feel manipulated, trust drops and refunds can rise.

Use holdout groups when you are making many changes

If you run frequent tests or large scale merchandising changes, keep a small holdout group that experiences the standard store. This helps you detect whether broader factors are affecting performance, not just the changes you are testing.

Account for seasonality and promotions

Pricing sensitivity changes during:

• Holiday periods
• Payday cycles
• High promotional seasons
• Weather-driven demand (for some categories)

If you must run a pricing test during a promotion, aim for symmetry:

• Ensure both variants are equally eligible for the promotion.
• Keep email and ad traffic split consistent across variants.
• Use RPV or GPV rather than conversion rate alone, since promotions affect basket behaviour.

Common pitfalls in Shopify price testing (and how to avoid them)

Changing price breaks your attribution

Some analytics set ups record revenue differently depending on discounts or taxes. Validate that revenue is tracked consistently across variants. If you use Google Analytics or server-side tracking, confirm that the revenue and item price fields are passed correctly.

Variant contamination via discount codes

If Variant B is “lower price via code”, customers may share the code. That exposes control users to treatment pricing, diluting the effect. Solutions:

• Use automatic discounts rather than codes where possible.
• Use unique codes per user session if your tooling supports it.
• Exclude the tested product from broad codes temporarily.

Price consistency issues across international customers

If you sell internationally, currency conversion and rounding can create unintended price points. Confirm the exact displayed price per market and ensure your variants are meaningful in each currency. Consider limiting the test to one market if necessary.

Stockouts create false demand signals

If one variant sells more and causes inventory pressure, you might see conversion drops due to low stock messaging or longer delivery promises. Either ensure sufficient inventory or stop the test if inventory becomes a confounder.

Interpreting a non-significant result as “no effect”

A non-significant result can mean:

• There is truly no meaningful difference in outcomes
• The effect exists but is smaller than your sample size can detect
• Your test was noisy due to discounts, seasonality, or tracking issues

If the confidence interval includes both meaningful positive and negative outcomes, the correct conclusion is “inconclusive”. The next action might be a larger price step, a longer test, or a cleaner environment.

What to do after you find a winning price

Pricing optimisation is not a one-off project. After you identify a winning price point, build a simple programme:

• Lock in the winner and monitor for two to four weeks.
• Run a follow up test to check if there is additional headroom (for example, test +10% again if the first increase won strongly).
• Test value reinforcement: if a higher price won narrowly, improve messaging and test whether you can hold conversion while increasing margin.
• Expand to similar products: apply learnings by category, then validate with tests on other SKUs.

Link your tests to business constraints:

• If paid acquisition is limited by margin, optimise for profit per visitor.
• If your product is subscription oriented, consider first order profit separately from long term value.
• If your brand relies on premium positioning, include guardrails around review ratings and returns.

Where ConvertLab fits in a pricing workflow

Many merchants discover that price does not operate alone. A price that feels “too high” may become acceptable when the title clarifies who it is for, when the description addresses key objections, or when the page highlights warranty and delivery clarity.

ConvertLab is built for product page experimentation, which makes it practical to:

• Run a/b test prices shopify experiments on product pages without rebuilding themes
• Test price alongside titles and descriptions when you suspect perceived value is the constraint
• Evaluate outcomes using revenue-aware metrics so you do not over-optimise for conversion rate alone

It should still be used with good experimentation hygiene: clear hypotheses, clean tracking, defined run times, and guardrails based on refunds and customer satisfaction.

Conclusion and next steps

Price testing is one of the most direct ways to improve ecommerce profitability, but it requires disciplined experimentation. Focus on revenue or profit per visitor, choose meaningful price variants, run tests long enough to cover real buying behaviour, and keep the customer journey consistent from product page to checkout.

Next steps:

• Pick one high-traffic product and record a baseline of RPV, conversion, AOV, and refunds.
• Select two price points with a meaningful difference and define your guardrails.
• Run a clean A/B test for at least one full weekly cycle, then analyse results and roll out carefully.
• Repeat with follow up tests to map demand sensitivity and find your sustainable optimum.

CTA: Test prices and copy together

ConvertLab lets you test prices alongside your copy. Find the combination of title, description, AND price that maximises revenue per visitor; not just conversion rate.

Install ConvertLab from the Shopify App Store and run a controlled price experiment on a product page, then iterate based on what your customers actually do.

If you want deeper background on pricing experimentation strategy, see the pillar page: /convertlab/guides/price-testing.