Beating price wars and big players in the homogeneous product market

TL;DR

Difficulty: Intermediate

  • Identical products sold only on price end at zero margin. Bertrand's model of two sellers undercutting each other shows why, and nothing since has overturned it.
  • A price cut comes off the contribution, not the price. A 10% cut on a 25% margin needs 67% more units just to stand still, and a match does not bring them.
  • A supplier may recommend a price but may not enforce one. In the UK that is resale price maintenance and the CMA fines it.
  • The levers that work change what the customer compares: switching costs, visible service, bundles and baskets, and on Amazon the landed price including delivery.

If you sell the same branded products as the national chains, a price cut from one of them feels like a decision you cannot win. In this article I take the economics of price wars, from Bertrand's undercutting game to Porter's supplier power and Klemperer's switching costs, and turn it into the arithmetic a small seller can actually do, with an invented Leeds tyre retailer working through a rival's 10% cut three ways.

Selling the same product as everyone else is the hardest game in retail

If you sell branded tyres, branded electronics or branded pet food, your customer can buy the identical item from a national chain, from a marketplace and from three other shops in the same postcode. Economists call this a homogeneous product market. The thing in the box is the same whoever sells it, so the only question the customer seems to have left is "who is the least expensive?"

That question is a trap, and this article is about how to stop walking into it. The economics of price wars is old and well understood, and it says something uncomfortable and useful: when identical products compete only on price, the margin goes to zero, for everyone. The way out is not to be braver about cutting. It is to change what the customer is comparing.

Throughout I use an invented business, Treadline Tyres, an online tyre retailer in Leeds run by a man called Dev. He sells, fits and disposes of the same branded tyres the national chains sell. Nobody real, every number made up so that I can show every step. The one thing I ask you to read first is price elasticity in plain English, because elasticity decides most of what follows.

The textbook says a price war ends at zero margin, and the textbook is right

Start with the simplest model. Perfect competition, in the textbook definition, is a market where many firms produce identical products and no single firm can move the price, so every firm is a price taker, and any economic profit attracts new entrants until profit is driven down to zero.1 That is the long run story with many sellers.

The short run story with only two sellers is nastier, and it carries Joseph Bertrand's name. Two firms sell the same product at the same unit cost and set prices at the same time. Whoever is a penny lower takes the whole market. So each undercuts the other until neither can go lower without losing money, and the only stable outcome is both pricing at marginal cost. Two firms, zero profit, the same result as perfect competition. The industrial organisation textbooks call this the Bertrand paradox.2

The paradox only breaks when one of the assumptions breaks. If the products differ in the customer's eyes, if capacity is limited, if switching costs money or effort, or if the firms meet again next month, the price stays above cost. Keep that list. It is the whole strategy of this article.

The oligopoly chapter adds the piece that explains the sinking feeling when a rival cuts: rivals tend to match price cuts but not price rises, which is why a cut is so quickly answered and so rarely reversed.3

Worked example: two Leeds tyre sellers and a penny

Dev sells a mid range 205/55 R16 tyre fitted for £80. The tyre costs him £52 from the distributor, and fitting, valve, balancing and disposal cost him another £8, so his variable cost is £60 and his contribution is £20 a tyre. A rival across the city sells the identical tyre at £80 with the same costs.

The rival goes to £79. Every customer who compares now buys from the rival, so Dev goes to £78. The rival goes to £77. Nine rounds later both are at £61, making £1 a tyre, and a tenth round takes them to £60, where neither makes anything. Nothing in the product stopped this. The only things that could have stopped it are the things that make the two offers different.

Elasticity decides whether a price cut pays, and in a homogeneous market it rarely does for contribution

Price elasticity of demand is the percentage change in quantity demanded divided by the percentage change in price. Above one in size, demand is elastic, below one inelastic, exactly one unitary.4 In a homogeneous market the elasticity a single seller faces is high, because the customer loses nothing by walking to the rival. That makes revenue look encouraging after a cut, which is where the damage hides.

The revenue effect of a price change, given the elasticity, is:

R1R0=(1+Δp) (1+ε Δp)\frac{R_1}{R_0} = (1 + \Delta p)\,(1 + \varepsilon\,\Delta p)

Here R0R_0 is revenue before the change and R1R_1 after, Δp\Delta p is the proportional price change (a 5% cut is −0.05-0.05), and ε\varepsilon is the price elasticity of demand, a negative number because quantity moves against price. The first bracket is the price effect, the second is the quantity effect, and the product tells you whether revenue rose. The rule it encodes is the one the textbook gives: with elastic demand a price cut raises total revenue, with inelastic demand it lowers it.5 The definitions and the midpoint version are in Foundations: elasticity.

Worked example: Dev's spring sale

Last spring Dev trimmed the £80 tyre to £76, a 5% cut, and sold 675 instead of 600, a 12.5% rise. His elasticity on that line is therefore ε=12.5%−5%=−2.5\varepsilon = \frac{12.5\%}{-5\%} = -2.5, elastic. Revenue went from 600×80=£48,000600 \times 80 = £48{,}000 to 675×76=£51,300675 \times 76 = £51{,}300, and the formula agrees: (1−0.05)(1+2.5×0.05)=0.95×1.125=1.069(1 - 0.05)(1 + 2.5 \times 0.05) = 0.95 \times 1.125 = 1.069, up 6.9%.

So far it looks like a win. Now the contribution. Before the cut, 600×(80−60)=£12,000600 \times (80 - 60) = £12{,}000. After it, 675×(76−60)=£10,800675 \times (76 - 60) = £10{,}800. Revenue up £3,300, profit before fixed costs down £1,200. The cut gave away £4 on every one of 675 tyres, £2,700, and the extra 75 tyres brought in £16 each, £1,200. The sale was busy and it lost money.

A match costs more margin than it looks, so work out the volume you would need

Contribution margin is the amount by which a product's selling price exceeds its variable cost per unit.6 When you match a rival's cut, the whole cut comes off the contribution, not off the price, and the contribution is the smaller number. The question is never "can I afford to be 10% lower?" It is "how many more units do I need to sell at the lower contribution to end up where I was?"

q1q0=p0−cp1−c\frac{q_1}{q_0} = \frac{p_0 - c}{p_1 - c}

Here q0q_0 and q1q_1 are the units sold before and after the cut, p0p_0 and p1p_1 the prices before and after, and cc the variable cost per unit. The right hand side is the old contribution per unit over the new one. If the ratio is 1.67, you need two thirds more units just to stand still. This is marginal thinking: the next unit's contribution, not the average margin on the price list, decides whether a cut pays, and Foundations: marginal versus average covers why the average misleads.

There is a shortcut I use in meetings. Call the contribution margin as a share of price mm and the cut as a share of price dd. The volume you need rises by dm−d\frac{d}{m - d}, so the elasticity you would need just to break even is ε∗=−1m−d\varepsilon^* = -\frac{1}{m - d}. For Dev, with m=0.25m = 0.25 and a 10% cut, that is −10.15=−6.7-\frac{1}{0.15} = -6.7. His measured elasticity is −2.5-2.5. He is nowhere near.

Worked example: the 10% match on paper

The rival national chain cuts the 205/55 R16 to £72, a 10% cut. If Dev matches, his contribution per tyre falls from £20 to £12, and the volume ratio he needs is 80−6072−60=2012=1.67\frac{80 - 60}{72 - 60} = \frac{20}{12} = 1.67. To keep £12,000 a month he would need 1,000 tyres instead of 600.

A match does not do that. A match stops his existing customers leaving. It does not make the chain's customers drive to Leeds. So the realistic outcome of a match is 600 tyres at £12, which is £7,200, a loss of £4,800 a month against the month before, with the shop exactly as busy as it was.

Worked example: Treadline Tyres against the national chain, three ways

Now the full comparison, with Dev's assumptions written down so they can be argued with. Baseline: 600 tyres a month at £80, contribution £20, total £12,000. The chain is at £72.

Option one, full match at £72. Dev expects to hold 600 tyres. Contribution 600×12=£7,200600 \times 12 = £7{,}200.

Option two, no match, stay at £80. Dev is unsure how many customers will leave. He thinks there is a 50% chance he loses 30% of the volume, because many of his customers come back out of habit, and a 50% chance he loses 60%, because the comparison sites will show the gap to everyone. The expected volume is 600×(0.5×0.70+0.5×0.40)=600×0.55=330600 \times (0.5 \times 0.70 + 0.5 \times 0.40) = 600 \times 0.55 = 330 tyres, where 0.700.70 and 0.400.40 are the shares of volume kept in each case and 0.50.5 is the probability of each. Contribution 330×20=£6,600330 \times 20 = £6{,}600. Writing the scenarios down rather than picking one is Foundations: expected value in use, and it keeps two people from arguing about a single guess.

Option three, partial match with a service bundle. Dev goes to £76 and adds two things to every tyre: puncture repair for the life of the tyre and a free alignment check with the set. His records say one tyre in eight comes back for a repair that costs him £20, so the promise costs an expected 18×20=£2.50\frac{1}{8} \times 20 = £2.50 a tyre, and the alignment check is already in the fitting time. Contribution per tyre 76−60−2.50=£13.5076 - 60 - 2.50 = £13.50. He expects to lose 10% of volume to the chain's lower headline, so 540 tyres. Contribution 540×13.50=£7,290540 \times 13.50 = £7{,}290.

Four bars. Before the rival's cut, contribution was £12,000 a month. Not matching and staying at £80 gives an expected £6,600. Matching fully at £72 gives £7,200. A partial match at £76 with a puncture repair and alignment bundle gives £7,290. None of the three responses gets back to the original figure. Tap or click to open full screen and move around

The bars are monthly contribution on the one tyre line, in pounds, and the four labels are the month before the cut and Dev's three responses. Two things to take from it. First, none of the responses gets back to £12,000. A price cut by a bigger rival costs you money whatever you do, and the job is to choose the smallest loss, not to find a clever escape. Second, the differences between the three responses are a few hundred pounds a month, which is inside the error of Dev's guesses. The month's figures alone do not pick a winner.

What does pick one is the second purchase, which the month's figures cannot see. Dev's records say a customer comes back for the next set after about 30 months. Customers who have used his puncture promise return 45% of the time, customers without it 25%. At two tyres and £20 each, the expected contribution from a returning customer is 0.45×40=£180.45 \times 40 = £18 with the bundle and 0.25×40=£100.25 \times 40 = £10 without, where 4040 is the contribution on a pair of tyres and the decimals are the return rates. On 270 bundle customers that is £4,860 of later contribution against £3,000 from 300 matched customers. The bundle is not a nicer discount. It is a switching cost, which is the subject of the section after next. The longer version of the repeat purchase arithmetic is in the retention maths.

Suppliers hold more power than you think, and the law limits what they can do with it

The old version of this article described a supplier being leaned on by the big resellers to keep a small one out, and then being asked to impose one price on everyone. The first half of that is ordinary supplier power. The second half is illegal in the UK, and I should have said so.

Michael Porter's five forces framework puts suppliers next to rivals, customers, new entrants and substitutes as a source of pressure on profit, and it names the conditions under which rivalry turns destructive: when it gravitates to price alone, when products are nearly identical and when buyers face few switching costs.7 A tyre distributor whose largest accounts are the national chains has every commercial reason to serve them first. Credit terms, allocation in a shortage and marketing support will follow the volume. In my experience a small reseller gets a better hearing from a distributor by showing a clean payment record and a sensible forecast than by promising rapid growth, because the distributor has heard the growth story before.

What the distributor may not do is fix your price. The Competition and Markets Authority (CMA) is explicit: a supplier must not dictate the price at which its products are sold, online or anywhere else, a policy that sets a minimum advertised price for online sales can equate to resale price maintenance and is usually illegal, and a supplier must not use threats, financial incentives, withheld supply or worse terms to make resellers stick to a recommended price.8 Recommending a price is generally lawful. Enforcing it is not. The same letter warns resellers that if you agree to sell at a fixed or minimum price, you may both be breaking the law, and that fines can reach 10% of a business's worldwide turnover.9 The CMA's examples are not exotic: light fittings, commercial refrigeration, bathroom fittings, musical instruments, all fined for policing online prices.

So the "uniform pricing" the old article described as a competitive setback is in fact a complaint to the CMA. If a supplier tells Dev he may not go below a figure, the right response is a polite email asking for that in writing, and then a conversation with a solicitor. The rules on watching rivals' prices are a separate topic and I cover them in competitor price monitoring in the EU.

Worked example: the distributor's phone call

Dev's distributor rings to say the chain has complained about his £76 price and asks him to "respect the recommended retail price" (the RRP) of £80 or risk losing next season's allocation. Dev replies that he sets his own prices, asks the distributor to put the request in writing, and keeps a note of the call with the date. The distributor does not put it in writing. Dev has his allocation and his price. Had he agreed, both of them would have been at risk, and the chain would have got the price floor it wanted at no cost to itself.

Switching costs turn identical products into different ones

Paul Klemperer's 1987 paper makes the point that matters most for a small seller: products that are identical before a purchase become differentiated after it, because the customer now faces a cost to switch. Learning costs, transaction costs and costs the firm creates on purpose, such as repeat purchase discounts, all count.10 Once a customer has a reason to come back to you that the rival cannot copy by changing a price, the Bertrand undercutting stops working on that customer.

Klemperer adds something that explains the chain's behaviour. Where switching costs exist, firms compete hard for market share early, because a customer won now is worth a stream of later purchases.10 The chain's 10% cut may not be about this month's tyres at all. It may be buying customers to keep.

Worked example: Dev's tyre record

Dev builds three small things into his shop. Every customer gets a tyre record: sizes, pressures, fitting dates and tread depth at the last visit, with a reminder six weeks before the MOT is due. Every tyre carries the puncture promise from the worked example above. Every returning customer sees their previous order and can reorder the same tyre in one tap. None of these costs the customer anything, and all three cost the customer something to give up. The identical tyre is now a different offer, and the chain's headline price has to beat the whole offer rather than the number.

Service is the differentiation left when the product is not

Theodore Levitt wrote in 1980 that there is no such thing as a commodity, and that all goods and services are differentiable.11 His point was not that you can pretend a tyre is special. It was that the product is only part of what the customer buys. The rest is how it is found, bought, delivered, fitted, explained and supported, and every one of those can be done better or worse.

The discipline is to pick service elements that the customer values more than they cost you, and to make them visible before the price comparison happens, not after. A benefit the customer only discovers at the fitting bay cannot stop them clicking the rival's lower number.

Worked example: four things Dev makes visible

Dev adds four lines above the price on every product page: a fitting slot you choose, including Saturday mornings, a photo of your old tyre's tread sent to your phone before it is taken off, the price held for 48 hours once you have picked a slot, and the puncture promise. Each costs him minutes, not pounds. On the comparison site his tyre still shows £76 against the chain's £72. On his own page a customer who has got that far sees £4 against four things the chain does not list. That is the comparison he wants.

Loss leaders and bundles move the fight from a product to a basket

Retailers have always known that some prices exist to bring people in. Rajiv Lal and Carmen Matutes modelled it in 1994: a store advertises some products below cost to draw shoppers who then buy the rest of their list at full margin, and the products most likely to be priced that way are the ones customers value least on their own.12 Bundling is the other side of the same coin. William Adams and Janet Yellen showed in 1976 that selling goods together can earn more than selling them apart when customers value the parts differently.13

For a seller of identical products the lesson is sharp. If the chain's £72 is a loss leader, matching it on the tyre alone is fighting on the ground the chain chose. The right unit of comparison is the basket.

Worked example: the alignment attach rate

Dev tries the match at £72 but on his own terms. Alongside the tyre he offers four wheel alignment at £45, with a contribution of £35, and prices the pair at £110 as a bundle. His records say 40% of customers who see the bundle take it. His expected contribution per tyre sale is now 12+0.4×35=£2612 + 0.4 \times 35 = £26, where 1212 is the matched tyre contribution, 0.40.4 the attach rate and 3535 the alignment contribution. That beats the £20 he made before the cut. The tyre is a loss leader for the alignment, and the chain, if it is doing the same, is the reason it could afford £72 in the first place. The caveat is in the arithmetic: at a 20% attach rate the figure is £19, below where he started. The basket has to exist before the loss leader is anything other than a loss. How to design the offer at the checkout is a topic of its own in cross sell and upsell at checkout.

A price match guarantee is not the bargain it looks like

The obvious move in a homogeneous market is to promise to match any price. It feels like the smallest possible cut, because you only pay it when asked. The evidence says something odder. A study of advertised tyre prices found that a retailer's own guarantee had no significant effect on its own advertised price, but when more firms in a market advertised guarantees, prices rose, by roughly 10% of the average advertised price where every firm had one compared with none.14 A 2021 structural study of the tyre market found that price matching guarantees keep transaction prices between 1% and 8% higher than they would otherwise be, with the strongest effect on the most price sensitive customers.15

The economics is the Bertrand game in reverse. If every seller promises to match, no seller gains anything by cutting, so nobody cuts, and prices settle higher. That is not a reason to adopt one, because it is also the reason competition authorities look at them. It is a reason not to read a rival's guarantee as weakness. It is usually a signal that they would rather not fight at all.

Worked example: what the guarantee actually costs Dev

Dev considers "we match any Leeds price". On his elasticity the customers who would actually invoke it are a minority who have already found the chain's £72. If one customer in ten does, his average price falls by 0.1×8=£0.800.1 \times 8 = £0.80, where 88 is the gap and 0.10.1 the share who ask, and his contribution per tyre from £20 to £19.20. That is far less than the £8 a blanket match costs. But he has also told the chain that any cut it makes will be met, which, if the chain reads the same papers, is a reason for it to stop cutting. He puts the guarantee on the page, leaves his price at £80, and treats it as the defensive move it is.

Many homogeneous product sellers meet the price war on a marketplace, where the product page is shared and only one seller gets the main add to basket button. Amazon calls that the Featured Offer, and most sellers still call it the Buy Box. Amazon's own selling guidance for the UK lists what it weighs: only Professional plan sellers are eligible, new and used items compete separately, the landed price including delivery is compared across offers, delivery speed and reliability count, order defect rate and other service metrics are taken into account, and the product must be in stock.16

Price is one factor among six, and the price that counts is the price including delivery. Reliability, stock and account health are gates. Being the lowest headline price with a late delivery record does not win, and being £2 higher with faster, more reliable delivery can. It is Levitt's point enforced by an algorithm.

Worked example: landed price

Dev lists the tyre on Amazon at £78 with £6 delivery. The chain lists at £80 with free delivery. The landed prices are £84 and £80, so Dev is the more expensive seller despite the lower headline. He moves to £80 with free delivery, matches the landed price, and his offer is now judged on delivery speed and his defect rate, both of which he controls. He has not cut his price at all. He has moved the comparison onto ground where a small, careful seller can win.

The decision, step by step

Pulling it together, here is the order I work through when a bigger rival cuts the price of a product we both sell.

Tap or click to open full screen and move around

Reading from the top: the first question is Levitt's and Klemperer's, whether the customer is really comparing identical things. If not, the price is not the problem and should not be the answer. The second question is the contribution arithmetic from the middle of this article, which needs an elasticity measured from your own history rather than borrowed from a benchmark. The third question is Lal and Matutes: whether there is a basket or a service that makes the match pay. The last box is there because every one of these numbers was a guess until the month is over.

What I would do on Monday

The old version of this article ended with patience, innovation and a referral scheme. Patience is still right, and referral is a real lever I write about in growth marketing through retention and referral, but neither is the first thing to do when a rival cuts. The first thing is arithmetic. Take the product's contribution per unit, not its margin on the price list. Take the elasticity from your own last price change. Work out the volume a match would need, and ask honestly whether a match brings that volume or merely stops a loss. Then look at what sits around the product, the fitting, the delivery, the record, the promise, and ask what the customer would give up by leaving.

The textbook is not wrong. Identical products sold only on price end at zero margin. The whole of the small seller's job is to make sure the products are not identical by the time the customer compares them.

If a bigger rival has just cut and you would like the match, no match and bundle arithmetic done on your own numbers, I run a free workshop and the first two weeks of work are on me. Bring the price list and last year's volumes, I will bring the elasticity.

Sources

  1. 1.

    OpenStax, Principles of Microeconomics 3e, section 8.1 Perfect Competition and Why It Matters, Rice University, 2022, https://openstax.org/books/principles-microeconomics-3e/pages/8-1-perfect-competition-and-why-it-matters

  2. 2.

    Chrysovalantou Milliou, Static Oligopoly, section 3.3 Price Competition, lecture slides based on Paul Belleflamme and Martin Peitz, Industrial Organization: Markets and Strategies, second edition, Cambridge University Press, Athens University of Economics and Business, https://eclass.aueb.gr/modules/document/file.php/ERASMUS151/LECTURES/3bStaticOligopolyBNEW.pdf

  3. 3.

    OpenStax, Principles of Microeconomics 3e, section 10.2 Oligopoly, Rice University, 2022, https://openstax.org/books/principles-microeconomics-3e/pages/10-2-oligopoly

  4. 4.

    OpenStax, Principles of Microeconomics 3e, section 5.1 Price Elasticity of Demand and Price Elasticity of Supply, Rice University, 2022, https://openstax.org/books/principles-microeconomics-3e/pages/5-1-price-elasticity-of-demand-and-price-elasticity-of-supply

  5. 5.

    OpenStax, Principles of Microeconomics 3e, section 5.3 Elasticity and Pricing, Rice University, 2022, https://openstax.org/books/principles-microeconomics-3e/pages/5-3-elasticity-and-pricing

  6. 6.

    OpenStax, Principles of Accounting, Volume 2: Managerial Accounting, section 3.1 Explain Contribution Margin and Calculate Contribution Margin per Unit, Contribution Margin Ratio, and Total Contribution Margin, Rice University, 2019, https://openstax.org/books/principles-managerial-accounting/pages/3-1-explain-contribution-margin-and-calculate-contribution-margin-per-unit-contribution-margin-ratio-and-total-contribution-margin

  7. 7.

    Michael E. Porter, The Five Competitive Forces That Shape Strategy, Harvard Business Review, January 2008, https://hbr.org/2008/01/the-five-competitive-forces-that-shape-strategy

  8. 8.

    Ann Pope, Restricting resale prices: an open letter to suppliers and resellers, Competition and Markets Authority, 20 June 2017, https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/620454/resale-price-maintenance-open-letter.pdf

  9. 9.

    Competition and Markets Authority, Resale price maintenance: advice for retailers, GOV.UK, 21 June 2016, updated 29 June 2020, https://www.gov.uk/government/publications/resale-price-maintenance-advice-for-retailers

  10. 10.

    Paul Klemperer, Markets with Consumer Switching Costs, The Quarterly Journal of Economics, volume 102, issue 2, pages 375 to 394, Oxford University Press, 1987, https://econpapers.repec.org/RePEc:oup:qjecon:v:102:y:1987:i:2:p:375-394.

  11. 11.

    Theodore Levitt, Marketing Success Through Differentiation, of Anything, Harvard Business Review, January 1980, https://hbr.org/1980/01/marketing-success-through-differentiation-of-anything

  12. 12.

    Rajiv Lal and Carmen Matutes, Retail Pricing and Advertising Strategies, The Journal of Business, volume 67, issue 3, pages 345 to 370, University of Chicago Press, 1994, https://ideas.repec.org/a/ucp/jnlbus/v67y1994i3p345-70.html

  13. 13.

    William James Adams and Janet L. Yellen, Commodity Bundling and the Burden of Monopoly, The Quarterly Journal of Economics, volume 90, issue 3, pages 475 to 498, Oxford University Press, 1976, https://ideas.repec.org/a/oup/qjecon/v90y1976i3p475-498..html

  14. 14.

    Maria Arbatskaya, Morten Hviid and Greg Shaffer, Promises to match or beat the competition: evidence from retail tire prices, Advances in Applied Microeconomics, volume 8, pages 123 to 138, 1999, https://ueaeprints.uea.ac.uk/10809/

  15. 15.

    Samir Mamadehussene, Measuring the competition effects of price-matching guarantees, Quantitative Marketing and Economics, volume 19, pages 261 to 287, Springer, 2021, https://ciencia.ucp.pt/en/publications/measuring-the-competition-effects-of-price-matching-guarantees/

  16. 16.

    Amazon, Pricing and fees for selling on Amazon, frequently asked questions, How does the Amazon Buy Box work and how do sellers become eligible, Amazon UK, 2026, https://sell.amazon.co.uk/pricing