A product can earn more revenue even when fewer shoppers buy it. The reverse is also true: a lower price can lift the conversion rate without generating enough revenue to justify the margin you give up. Pricing decisions need customer behavior, not internal consensus or competitor benchmarks. The lesson is not to raise or lower prices, but it is to test your store's demand.
For the foundation, see how product price testing works on Shopify.
What is Product Price Testing?
Product price testing is an A/B experiment that shows separate visitor groups two base prices for the same product and compares revenue and conversion rate. Unlike discount testing, it changes the base price rather than adding a coupon, promotion, or shipping offer.
In GemX, Control A uses the current price, and Variant B uses one proposed price. Traffic is divided between them, while the assigned price remains consistent across sessions and future visits. The goal is to compare revenue outcomes, subject to a margin check, rather than select the highest conversion rate.
Learn more: How Product Price Testing Works on Shopify
7 Product Price Tests You Can Adapt to Your Shopify
Each idea below addresses a different pricing question, but the experiment structure stays the same. You will define the business context, write one hypothesis, compare the current price with one Variant price, and decide how revenue and conversion rate will guide the decision.
#1. Test a Higher Price on a Best-selling Product
A best-selling product gives you steady traffic and a known purchase pattern. Strong demand may mean that its current price leaves revenue on the table, especially if you set that price before the product had a proven market.
Hypothesis: Your test should state the expected trade-off before traffic enters the experiment:
"If we raise the product price from $X to $Y, revenue will increase because demand is strong enough to absorb the change without a disproportionate drop in conversion rate."
How to set up in GemX:
-
Control A should use the current live price
-
Variant B uses one realistic, higher price
-
A 50/50 traffic split is a practical starting point when traffic and business risk allow it
-
Keep the title, images, description, and options identical so the experiment isolates the price change
Real-world evidence:
The Original Grain generated the most profit at $799, although the team had assumed customers would not pay more than $500. The case study also reports 40% higher profit per session across a broader initiative involving ten top-selling products. That portfolio-level figure was not the result of the single $799 test.
How to read your result:
Revenue and conversion rate need to be read together.
-
If revenue rises while conversion rate slips only slightly, the higher price may be the better option if your margin check supports the decision.
-
If both revenue and conversion rate fall, the current price is the stronger choice for this product and test period.
-
If revenue remains flat while conversion rate falls, the current price is usually safer unless the higher price materially improves margin, which you should verify outside GemX.
Next move: If the higher price wins, use it in a separate follow-up test with a smaller price step. You can prioritize high-impact CRO experiments instead of moving directly to a much larger increase that introduces unnecessary risk.
#2. Test Whether a Lower Base Price Unlocks Enough Demand
This test suits a product that attracts qualified traffic but converts below expectations. Reviews, customer surveys, support conversations, or competitive research may indicate price sensitivity. The question is whether additional buyers can compensate for the revenue and margin surrendered on each unit.
Hypothesis: State the required outcome plainly: “If we lower the base price from $X to $Y, the increase in conversion rate will be large enough to raise total revenue." This wording prevents a conversion lift from becoming an automatic win.
How to set up in GemX:
-
Control A should retain the current base price
-
Variant B should use a lower base price
-
Keep the product content the same
This experiment compares two product prices, so it should not introduce a coupon, sale badge, countdown, or other promotional message that could influence the result.
VKTRY Gear increased conversion rate by more than 30% and profit per visitor by 14% after testing a price decrease. A lower price generated sufficient additional demand in this case, but it is not a target lift for another store.
How to read your result:
A higher conversion rate matters only when the resulting revenue supports the economics of the lower price.
-
If both metrics improve, calculate the unit margin outside GemX before adopting the Variant.
-
If the conversion rate rises while revenue stays flat or falls, the extra orders may not offset the lower revenue per unit.
-
A result can therefore look strong in the conversion column and still fail the business case.
Next move: If the lower price wins and the margin remains acceptable, run a separate test with a smaller decrease between the original and winning prices. That follow-up can show whether you can retain most of the demand lift while recovering some margin.
#3. Test a Competitor-aligned Price Instead of Copying It
A premium-priced product often sits beside lower-priced substitutes in search results and comparison pages. Competitor pricing can reveal a useful test range, but a spreadsheet cannot measure how your visitors value your brand, product quality, service, or delivery experience.
Hypothesis: Frame the benchmark as a question: “If we move from our premium price to a competitor-aligned price, conversion rate will increase enough to produce more revenue." The hypothesis makes the required revenue trade-off explicit.
How to set up in GemX:
-
Control A should use your current premium price
-
Variant B uses one price aligned with a genuinely comparable competitor.
-
Keep the positioning, images, description, and product options identical. Otherwise, the test would compare two different value propositions rather than two prices.
Real-world evidence:
Beards & Daisies positioned its products at a slight premium and questioned whether competitor-aligned pricing would improve performance. Lowering the price of its popular Peach Anthurium to match competitors produced 8.25% higher profit per visitor for the lower-priced option.
Other headline results in the case study cover the brand's broader testing program, so they should not be attributed to this individual price test.
How to read your result:
-
If the competitor-aligned Variant generates more revenue, the result shows price sensitivity for this product, audience, and test period. It does not prove that the entire brand should become cheaper.
-
If the current price wins, your visitors may value enough differentiation to support the premium, although the finding still applies only to the product and conditions tested.
Next move: Apply the learning to another comparable product only through a separate experiment. Catalog-wide repricing would assume that different products have the same demand curve, competitive set, and perceived value, which the first test cannot establish.
#4. Revisit the Price of a Newly Launched Product After It Gains Traction
A team may launch a product conservatively because it has little evidence of demand. Once the product is active, available through the Online Store channel, and receiving steady or unexpectedly strong traffic, the launch price becomes testable. This is an early post-launch experiment, not a pre-launch test.
Hypothesis: The test should connect stronger demand with a measurable trade-off: “If we raise the early launch price from $X to $Y, revenue will increase while conversion rate remains within an acceptable range because observed demand is stronger than expected.” Define that acceptable range before the experiment begins.
How to set up in GemX:
-
Control A should use the current launch price
-
Variant B should use one higher price
-
Keep the product title, images, description, and options identical between versions
Real-world evidence:
HAIRtamin launched its Miracle Brush conservatively before the product went viral on TikTok. The brand tested a small increase while traffic was strong and reported that conversion rate fell by around 2%, while revenue per visitor increased meaningfully. HAIRtamin adopted the higher price, increased media spend, and reached its December revenue target within 10 days.
How to read your result:
-
If the higher-priced Variant increases revenue while the conversion rate stays within the range you defined, the new price may better reflect the product's demonstrated demand. You should still check the unit economics outside GemX before adopting it.
-
If revenue falls or conversion rate drops beyond the acceptable range, the product's momentum may be more price-sensitive than traffic or social engagement suggested. The launch price remains the safer option until a different hypothesis warrants another test.
Next move: Revisit the decision after traffic settles into a more typical pattern. Viral or launch-period visitors may not represent the audience that will sustain the product over time.
Important note: Always start by building a testable pricing hypothesis before you set the Variant price.
#5. Test a Price Increase Before Passing Higher Costs to Customers
Higher supplier, production, freight, or tariff costs can compress product margin. A permanent increase may protect that margin, but it can also suppress demand. Testing one representative product first reveals the customer response before you change more of the catalog.
Hypothesis: The expected outcome should be commercially specific: “If we increase the product price from $X to $Y, revenue will hold or rise because customers value the product enough to absorb the increase.” The test does not need a conversion lift to support the higher price.
How to set up in GemX:
-
Control A should use the current price
-
Variant B should use one proposed cost-adjusted price
-
Keep every product element other than the base price unchanged
Minnow tested prices after tariffs increased its production costs. Shoplift reports that customers continued converting at their normal rate after a 10% increase on core products. The result supported protecting quality and margin through a higher price.
SAXX provides supporting evidence from a broader setup. A 5% increase across one collection produced no statistically significant conversion-rate difference after four weeks. SAXX expected a 1% to 2% gross-margin improvement over six months.
How to read your result:
-
If revenue holds and conversion rate remains stable, the result supports passing the tested cost increase to customers. Your margin calculation should confirm whether the gain solves the original problem.
-
If revenue rises despite a modest conversion-rate decline, the higher price may still be viable. If both metrics fall, consider a smaller increase or a separate response to costs instead of forcing the planned price through.
Next move: Test another product separately before applying the increase more widely. Different products can have different demand curves, even when the same supplier or tariff change affects them.
#6. Test a Round Price Against a .99 or .95 Ending
A one-cent or five-cent difference looks minor, but the ending may change how shoppers interpret the offer. A .99 price can signal value or promotion, while a round price may feel simpler or more premium. Neither signal is a universal winner.
Hypothesis: The direction should reflect your product's positioning. A value-led test might ask: “If we use a .99 ending, the conversion rate will increase because the product feels more affordable.”
A premium test might ask, "If we use a round price, revenue will hold because the price fits the product's positioning better.”
How to set up in GemX:
-
Control A and Variant B should use two actual base prices
-
Change only the base price and its ending
-
Keep the product content and presentation identical
-
Do not add a crossed-out compare-at price, discount badge, or promotional message to either version
Beards & Daisies reports comparing round prices with .99 and .95 endings, but the case study discloses neither a winner nor an uplift. The Journal of Retailing and Consumer Services also found that 9-ending effects differ by product category, brand positioning, and customer preference. The evidence supports testing, not declaring a default winner.
How to read your result:
-
If one ending increases revenue, use conversion rate to understand whether stronger demand or more revenue per order contributed to the outcome. Your separate margin check should confirm whether the difference is commercially worthwhile.
-
If revenue and conversion rate remain effectively flat, choose the ending that best fits your positioning and operating needs. A neutral outcome is still useful because it removes the need to keep debating a low-impact decision.
Next move: Avoid running multiple follow-up tests on minor price endings unless the product has enough traffic and the decision matters financially. A more substantial pricing question may deserve that traffic first.
#7. Challenge a Planned Increase and Be Ready To Keep the Current Price
A brand may assume that stronger awareness gives it room to charge more. The assumption is especially tempting when competitor benchmarks shaped the original price. A useful experiment must be allowed to reject the increase.
Hypothesis: The planned change should remain falsifiable: “If we raise the product price from $X to $Y, revenue will improve because customers' willingness to pay has grown with the brand.” Before launch, define what revenue and conversion-rate trade-off would justify adopting the Variant.
How to set up in GemX:
-
Control A should use the current product price
-
Variant B should use one proposed higher price
-
Keep the title, images, description, options, and other product information identical
-
Record the decision rule before launch so the team does not move the goalposts after seeing the data
Dossier launched with competitor-led price tiers of $29, $39, and $49 before testing those assumptions. Its published Intelligems case study says the price tests found little elasticity and that higher prices were not a path to greater profit per visitor.
The case study's 7.4% overall improvement in profit per visitor came from Dossier's broader experimentation program, not from a price increase.
How to read your result:
-
If the current price produces the stronger revenue outcome, keeping it prevents a costly rollout. The experiment succeeded because it resolved the business question, even though the proposed Variant did not win.
-
If the higher price wins under the decision rule you set, you have support for that product, audience, and test period. If neither version creates a clear, commercially meaningful difference, keep the current price until you have a stronger reason to test another change.
Next move: Investigate a different lever in a separate experiment instead of combining price with new content, shipping, or an offer. These other high-impact e-commerce A/B testing ideas can help you choose the next question without compromising the price test's interpretation.
Turn Test Results Into Your Next Shopify Price Experiment
A winning price should narrow your next question, not end the pricing work. One A/B result shows which of two prices performed better for that product, audience, and period, not a permanent optimum.
As GemX setup compares Control A with one Variant B. Therefore, a follow-up Shopify price experiment should begin as a separate two-price test rather than adding $45, $50, and $55 to one experiment.
For example:
-
Experiment 1: Compare a $50 Control with a $55 Variant.
-
Decision point: If $55 produces the stronger revenue outcome, treat it as the supported price for the next question.
-
Experiment 2: Compare a $55 Control with a new $59 Variant in a separate experiment.
-
Stop condition: If $59 reduces revenue or produces no commercially meaningful improvement, retain $55 as the last supported price.
Each experiment answers one directional pricing hypothesis and leaves a clear record of what the store tested. The sequence also makes it easier to stop when a new price no longer improves the business outcome.
This logic adapts the iterative principle behind the Double Down pattern. As this case uses three price groups, only the principle of narrowing around a previous winner carries over to GemX.
Before starting the follow-up, write a new hypothesis and decision rule, then review how to run an A/B test properly. A winning price remains a supported decision for a specific product, audience, and time. New traffic, costs, competitors, or customer expectations can justify another Shopify price experiment later.
How to Analyze Price A/B Test Results with Revenue and Conversion Rate
A price test needs revenue and conversion rate because the metrics answer different questions. The revenue result shows which version generated more sales value, while conversion rate explains how buyer response changed.
|
Revenue |
Conversion Rate |
Likely interpretation |
Next action |
|
Up |
Up |
Variant B generated more revenue and converted more visitors. |
Check margin and data stability, then consider adopting Variant B. |
|
Up |
Down |
Fewer visitors purchased, but the higher value from those purchases produced more revenue. |
Check margin outside GemX and decide whether the conversion trade-off is acceptable. |
|
Down |
Up |
More visitors purchased, but the additional orders did not compensate for the lower price. |
Keep Control or test a smaller decrease in a separate experiment. |
|
Flat or unclear |
Flat or unclear |
Neither version produced a decision-ready advantage. |
Keep Control and refine the pricing hypothesis before testing again. |
Important note: More revenue does not prove that profit also increased.
Before changing the live price, you should confirm that the pattern remains stable within the collected data. Use the decision rule written before launch, especially when revenue and conversion rate move in opposite directions. These product page performance metrics provide wider context, but the final decision should remain tied to the pricing hypothesis.
How to Match Your Pricing Strategies To the Right Experiment Type
Popular lists often mix base-price tests with broader e-commerce pricing strategies. While product price testing compares one Control price with one Variant one, it should not be used to represent these experiments:
-
Offer Structure: Bundles and volume discounts change what customers receive or how quantity affects price.
-
Promotion: Discount depth and compare-at-price visibility introduce mechanics beyond the base price.
-
Order Economics: Shipping rates and thresholds change the order cost rather than the product price.
-
Billing and Audience: Subscriptions and market-specific prices require different purchase or targeting logic.
-
Test Architecture: A three-price straddle exceeds the documented A/B comparison.
Pro tip: Use Product Price Testing to compare the current product price with one proposed price. Follow the guide to create a Product Price Testing experiment in GemX.
Final Thoughts
Your best first price test is not necessarily the cleverest psychological pricing idea. It is the unresolved pricing decision attached to an eligible, high-impact product where a wrong permanent change could affect meaningful revenue. If this pricing question is worth testing, install GemX and start comparing your prices today!
GemX: CRO & A/B Testing gives you the structure, metrics, and confidence needed to turn pricing decisions into proven growth without complex setup.
FAQs
