Competition-Based Pricing: Decide Without Giving Up Margin
Competition-based pricing compares your product with eligible competitor offers. Your team can price below, match, or sit above that reference. Change the SKU’s price only when the product match is reliable and the price clears the channel’s margin floor. Otherwise, keep the current price.
The essentials
- Competition-based pricing compares eligible offers. Use competitor prices as a reference. Move a price only after checking product-match quality and the SKU’s margin floor.
- In the gardening retailer’s 23,365-product catalog, only contested products (8.2%) needed routine price decisions. A catalog-wide price cut would have missed that split.
- Check product identity, availability, delivery, data freshness, and unit economics before an observed price enters any rule.
- Every automated rule needs a floor, a price corridor, a stop condition, an owner, and a review date.
- myPricing monitors offers, matches products, and can propose a price within your rules. Your team can change the proposal or stop the rule.
A family retailer in southwest France sells 25,000 gardening and pet-supplies products. After fifty years in business, it needed growth without lower prices or more advertising.
The gardening market had fallen 6.2%, while DIY was down 1.5%. Suppliers had raised purchase prices by 7 to 8% on the top 100 sellers. More than 20 merchants carried the same products.
Yet only one group among 23,365 classified SKUs needed a lower price. You can read the results in the full customer case.
Start with only truly comparable offers
Competition-based pricing uses other sellers’ offers as a reference. An eligible offer covers the same product, targets the same buyer, and makes a comparable delivery promise. After that check, you can price below, match, or sit above the market. Keep your price when the products differ or the market price falls below your profit floor.
In the European Commission’s e-commerce inquiry (2017), 53% of responding retailers tracked competitors’ prices. Of those trackers, 67% used software. Among software users, 78% adjusted prices: 43% manually, 27% with manual and automatic changes, and 8% fully automatically. Software gathers observations. Your team decides which ones deserve a response.
Use three pricing methods. Each answers a separate question:
| Method | Starting input | Answers well | Cannot decide alone |
|---|---|---|---|
| Competition-based pricing | Eligible competitor offers | Where your offer sits | Whether you can afford it |
| Cost-based pricing | Landed cost, target margin | Your lowest acceptable price | Whether shoppers accept it |
| Value-based pricing | Willingness to pay | What your difference is worth | Whether the order is profitable |
The cheapest offer does not always win. Le Chen, Alan Mislove, and Christo Wilson collected the first two pages of third-party offers for 1,000 bestselling Amazon products. They found 543 sellers using pricing algorithms.
Those sellers won the Buy Box more often at almost every price position. The Buy Box is Amazon’s featured purchase offer. Seller feedback and sales volume helped them win without always posting the lowest price (WWW '16, 2016).
Build a competitor set your team can explain
Include a seller only when a shopper could reasonably choose it instead of you. Match the product, the buyer’s market, and the delivery conditions. A private-label product has no exact rival, so start with cost and customer value. Use similar products only to gauge the price range of the category.
Check every offer before it enters a price calculation:
| Check | Pass | Refuse or investigate |
|---|---|---|
| Product identity | Same model, variant, bundle | Similar title, different contents |
| Condition and tax display | Same condition and tax display | New/used mix or inconsistent tax display |
| Offer eligibility | Available to the same buyer | Out of stock or stale listing |
| Delivery promise | Comparable shipping and lead time | Cheaper price, weaker promise |
| Seller context | Credible seller, similar service | Unknown seller, unusual returns |
| Data freshness | Recent, timestamped observation | No timestamp, erratic scrape |
| Unit economics | Candidate price clears your floor | Market price sits below your floor |
The chosen reference changes the answer. In the same Amazon study, 70% of sellers that followed the lowest price stayed within $1 of it. Among sellers that followed Amazon’s own price, only 40% stayed that close.
Next, verify product matching. This is the link between your SKU and the correct competitor listing. One motorcycle-parts distributor could not rely on EAN codes.
A brake cable fits a specific bike model, but no API exposes which cable fits which model. myPricing’s AI matching compares the image, title, and attributes, then scores the pair. The motorcycle-parts distributor configured these thresholds:
| Match confidence | Action | Who decides |
|---|---|---|
| Above 96% | Auto-approved | The engine |
| 90 to 96% | Weekly review queue | One team member |
The team reviewed those scores for six months before turning on automated pricing. The matching reached 99% accuracy across 200,000 SKUs. Profitability rose by 20 to 25%, while revenue rose 30%.
The price-monitoring guide explains how to collect offers. Monitoring gives your team an observed price. Your team decides whether to match it.
Make the margin floor a hard stop
Ask two questions for every product group. Are the eligible offers truly comparable? Does the proposed price stay above your floor?
Hold the price until the match is verified. For an eligible offer, check the floor before matching its price.
Keep the current price until your team verifies the product match. Reject any price below your floor.
Record the target position as a price index. Divide the candidate price by the reference price, with 1.00 meaning parity.
One home-equipment retailer, selling across more than ten marketplaces, builds its floor from daily supplier prices, marketplace commissions, logistics, shipping, returns, and fraud. It then adds a 10% minimum margin, raising that percentage in categories that can carry it. No repricing action can cross that floor.
This retailer chose 10%. Use the selling-price and margin calculator to find your own threshold.
The operator may accept a price above the line. Below the line, the operator rejects the match.
A loss leader is a product sold below its normal commercial floor to attract customers. Treat it as a controlled exception, not a permanent setting. Name the SKUs, budget, end date, and rollback condition before launch. Check local law too: several markets, including France, restrict or ban selling below cost.
Move from an observed offer to a published price
Your team should run three checks before changing a live price.
When a check fails, the system stops the update. The operator sees the reason for the refusal.
First, your team verifies the product and the offer’s availability. Your team then selects a defensible reference and compares the proposed price with the floor. If one check fails, the system stops the update and shows the reason.
Alberto Cavallo measured how often US retail prices changed each month. The rate rose from 15.4% in 2008-2010 to 27.4% in 2014-2017. That shift shortened the implied life of a price from 6.5 months to 3.7 months (NBER Working Paper 25138, 2018).
Reprice volatile SKUs continuously and review the others periodically. One high-tech seller uses AutoPilot on roughly 1,000 SKUs across two marketplaces. It puts about 30% of them on high-frequency repricing and reviews the other 70% periodically. Its reaction time fell from hours to minutes.
Automate one step at a time. Begin with monitoring, then add product matching and costs. Put guardrails around the proposed price before the system can publish it.
Dynamic repricing with margin protection executes the rule. Competitive intelligence matches products and tracks market position. Multi-channel pricing accounts for channel commissions.
myPricing proposes a price within your limits. You can change it or stop the rule.
Reactive rules can raise prices as well as lower them. Stephanie Assad, Robert Clark, Daniel Ershov, and Lei Xu studied German gas stations after pricing software spread in 2017. The adopting stations increased their margins by about 15%.
When every station in a local market used the software, the market margin rose about 38%. The researchers found no change in a duopoly with only one adopter (Journal of Political Economy, 2024).
Treat the four catalog states differently
The product has no direct rival. The retailer tests a higher price.
The retailer already wins with this product. It holds the price and watches conversion.
The retailer can reach a winning position. It tests a guarded price.
The retailer cannot compete on price. Its ads highlight free returns, seller ratings, and customer service.
In this catalog, the retailer had to routinely review the price of about one product in twelve.
The retailer split the catalog four ways. Products with no direct competitor accounted for 45.4% of the catalog. The retailer tested higher prices on those products when the market allowed it. Another 29.1% already held a winning position but needed visibility.
Those two groups received the advertising budget. The retailer then measured 34% more clicks and 30% higher conversion. Its advertising cost was 13% lower than for similar products without pricing intelligence.
Losing products made up 17.3% of the catalog. The retailer’s ads highlighted free returns, seller ratings, and customer service instead of a lower price. Only the contested products, 8.2% of the catalog, needed routine price decisions. Even there, a lower price was optional.
Set stop rules before automation
Write a stop condition beside every rule. Hold the SKU’s price when match confidence is too low. Ignore stale offers and reject prices below the channel floor.
Add a price corridor, which sets a minimum and maximum price. Cap reactions so one automated change cannot trigger another indefinitely.
When a cost changes, name the person responsible and the next review date. A competitor feed does not make that decision.
Clear these seven points before you switch a rule on
Complete this list for each product group:
- Name the pricing owner, approver, and next review date.
- Define direct, benchmark, and excluded competitors. Set the match-confidence threshold.
- List the eligibility fields, including stock and delivery promise.
- Choose the reference, such as low, median, high, or percentile. Set the target price index.
- Calculate landed cost and minimum margin. Set approval, automation, and product-hold rules.
- Record the cadence, stop trigger, rollback price, and KPI.
- Keep one auditable row per SKU and channel. Connect the triggering offer to the resulting margin.
Sources
- European Commission, E-commerce Sector Inquiry, COM(2017) 229 final and MEMO/17/1262, 2017.
- Le Chen, Alan Mislove, Christo Wilson, Algorithmic Pricing on Amazon Marketplace, WWW '16, 2016.
- Stephanie Assad, Robert Clark, Daniel Ershov, Lei Xu, Algorithmic Pricing and Competition, Journal of Political Economy 132(3), 2024.
- Alberto Cavallo, More Amazon Effects, NBER Working Paper 25138, 2018.
- Commission Regulation (EU) 2022/720 of 10 May 2022 on vertical agreements, Article 4, 2022.
- Anonymized Boostmyshop customer records: gardening and pet supplies, motorcycle parts, home equipment, and high-tech.
Frequently asked questions
Competition-based pricing sets your price against the eligible competitor offers a buyer can actually choose instead. You may price below, match, or sit above that reference. Product-match quality, channel costs, customer value, and your minimum margin decide whether the SKU moves or keeps its current price.
A gardening retailer classified 23,365 products. It found 45.4% with no direct competitor, 29.1% already winning, 17.3% unable to win on price, and 8.2% contested. Only the last group needed routine price decisions.
The method gives your team a market reference and exposes unusual prices. Check each offer, enforce a margin floor, and add a stop condition. Without those checks, your team may base its price on an incorrect product match, a stale offer, or an unprofitable competitor price.
You should not match it automatically. The lowest offer may be out of stock, exclude shipping, cover another variant, or fall below your margin floor. Compare eligible delivered offers first. Then choose to go below, match, stay above, or leave the price unchanged.
Set the cadence by product group. Cavallo measured US monthly price-change frequency at 15.4% in 2008-2010 and 27.4% in 2014-2017 (NBER, 2018). The high-tech seller’s 1,000-SKU catalog used high-frequency repricing for 30% and periodic checks for 70%.
You may monitor public prices, but you must not coordinate prices with competitors. Article 101 TFEU prohibits it in the EU. Set prices independently and follow the data source’s terms.
A supplier cannot impose it. Article 4 of Regulation (EU) 2022/720 treats resale price maintenance as a hardcore restriction and removes the agreement's block exemption. In the European Commission inquiry, two in five retailers still faced supplier pricing restrictions, and almost a third normally complied.
Continue exploring
What marketplaces read off your warehouse: availability, picking accuracy, dispatch time. Jonathan Coloma, COO of Repiauto, on going from 300 to 1,000 orders a day in two years: the paper wall, ERPs built for a different problem, a go-live quoted at fifteen days, and what changed in his working day.
To hold up under peak volume, split the order queue into waves based on real operating constraints, then decide where barcode checks belong. At the packing bench, the operator verifies the parcel. Carrier rules select the service and print the label, so the next order can start without another decision.
Industrializing your purchasing means ordering on rules, not guesswork: at the right time (the reorder point), in the right quantity (your coverage period), from the right supplier, with a purchase order validated in one click. Available on replay.
See it in your own operations
Book a 20-minute call. We'll map your workflow and point out where it breaks. No slides.
Free 20-min call · No commitment