What happens when growth carries the volume but not the margin?

Companies often add a product that grows faster than the business they started with, but makes much less money on each transaction. We looked at what happened next.

Thin margins can work. One-time transactions can work. The hard part is doing both.

The transaction is not the asset. What survives the transaction is.

The pattern

Low margin is not the problem. Nothing carrying forward is.

Plenty of good companies sell something at a margin that would look alarming on its own. Costco keeps merchandise margins near the floor. Robinhood does not charge for stock trades. Shopify earns far less on a payment than on a software subscription. Block makes almost nothing when a Cash App customer buys bitcoin.

In each case, the low-margin product is doing a different job. It brings in a customer who stays, or it earns a little more from a customer who was already there. The money shows up somewhere else.

Then there is the other group. Zillow bought houses and sold them. Groupon sold discounted merchandise. PayPal processed payments for large merchants through Braintree. These lines grew fast, and in two of the three cases they became the biggest thing on the income statement.

The difference is not how much money each transaction made. It is what was still there the morning after. A homeowner sells one house and does not come back for a decade. A merchant on Shopify sells something the next day, and the day after that.

That is the whole finding. When a company adds volume that leaves nothing useful behind, the volume tends to make the business bigger without making it stronger.

Where companies sit

Two questions, four positions

How much does each transaction earn, and does anything valuable stay behind when it is over. Almost every company in the study sorts cleanly on those two.

Something valuable carries forward

Little carries forward

Strong margins

Usually works

Costco
Title insurance
Vulcan Materials
Amazon advertising

The company earns enough on each sale and also keeps an advantage for the next one.

Can still work

Homebuilders such as NVR

The customer may never return, but the business can still work if each sale earns enough on its own.

Thin margins

Can work very well

Shopify Payments
Robinhood
Block / Cash App
McKesson
Netflix streaming

Some of the strongest businesses in the study live here on purpose. Thin margins are affordable when the relationship keeps paying.

Where we found the most trouble

Zillow Offers
Groupon Goods
PayPal / Braintree

We went looking for a long-term winner here and could not find a clear one. World Kinect came closest, but it did not provide a clear long-term success case.

If little carries forward, each transaction needs to be attractive on its own.

What happened

Seven companies, and what was left when the sale was over

Shopify

Shopify Payments earns about 38 cents of gross profit per dollar. The software subscription earns about 82 cents. On paper, growing payments makes the company look less profitable, and the blended gross margin has been falling for years.

But Shopify does not have to find a new customer for every payment. The merchant is already there, already paid for. Payments needs far less incremental sales and product cost than acquiring a new merchant. So while gross margin fell, operating costs dropped from 45 percent of revenue to 37, and gross profit grew 24 percent.

What carried forward: the merchant relationship, already paid for once.

Robinhood

Stock trading is free. In one recent quarter, $647 billion of stock changed hands and produced $86 million of revenue. Options, on a far smaller amount of trading, produced $304 million.

Robinhood has never tried to fix that ratio, because free trading is the point. The same customer leaves cash in the account, borrows on margin and subscribes to Gold. Cash sweep balances hit $35.4 billion and interest income passed options revenue.

What carried forward: the customer relationship and account balance.

Costco

Costco sells merchandise at margins that would sink a normal retailer. After store costs, the goods business earns very little.

Membership fees are under two percent of sales and roughly half of operating profit. Costco has run this way for forty years and has raised the fee rarely. The company sets that price, and the member either renews or leaves.

What carried forward: the membership, priced by Costco.

Zillow Offers

Zillow bought houses from homeowners, fixed them up and sold them. Because the whole sale price counted as revenue, the line grew enormously. In one quarter it produced $1.2 billion against $480 million for the core listings business.

The same quarter, the home-buying segment lost $422 million. The company wound it down and reduced headcount by about a quarter. A person sells you their house once. There was very little left afterwards that made the next purchase cheaper or easier.

What carried forward: very little.

Groupon Goods

Groupon's original business took a cut of local deals at very high margins. Then it started buying and selling physical merchandise, which grew to more than half of revenue at about 15 percent gross margin.

Leadership described the benefit as engagement and customer lifetime value. Those numbers were never published. Groupon exited the category in 2020, and in a later quarter revenue fell 30 percent while gross profit rose 13 percent. Most of the volume had not been worth much.

What carried forward: nothing anyone could measure.

Title insurance

A homebuyer might buy title insurance twice in a lifetime. By the customer test, this should be one of the worst businesses in the study. Pre-tax margins run around 15 to 17 percent.

Two things explain it. The buyer does not usually choose the provider. Lenders, agents and closing attorneys do, and more than half of Fidelity National Financial's premiums arrive through agents it works with continuously. Separately, every completed search adds property records to a searchable file, so the next search on that property starts from work already done.

What carried forward: the referral relationships, and a growing file of property information.

Netflix

This one is the important exception. At the end of 2011, DVDs by mail had 11.2 million subscribers and produced $194 million of profit. Streaming had 21.7 million subscribers and produced $52 million. Twice the customers, a quarter of the profit.

Anyone judging by margin alone would have concluded Netflix was moving customers into a worse business. What the margin was actually measuring was age. DVD costs rose with every disc mailed. Streaming costs were mostly fixed and had not yet been spread. Within four quarters the streaming margin had gone from 11 percent to 16, and then higher.

What carried forward: the subscriber, plus economics that improved as the base grew.

The common thread

Something survived the transaction

In every company that made low margins work, we could point to a specific thing that was still there when the sale was over. It was not always a customer.

The customer comes back

Costco, Robinhood, Netflix. The buyer returns, or leaves money behind, or keeps paying a fee.

The merchant comes back

Shopify. The next sale costs nothing to win because the relationship already exists.

The supplier or referrer stays

Title insurance, and the hotels behind a travel site. The person choosing the provider is not the person paying.

The data gets reused

Title files, Amazon's view of what shoppers buy. Past sales make future ones cheaper to serve.

The location advantage stays

Vulcan Materials. Crushed rock is too heavy to ship far, so the nearest quarry usually wins.

The cost advantage stays

Ryanair. When you are the cheapest operator by a wide margin, a thin price per seat is still a good business.

The hard cases

What happens when nothing carries forward

The companies that had the hardest time shared two traits. Each sale earned very little, and the next sale had to be won again from scratch.

Zillow bought a house from someone who would not need another for years. Groupon sold a discounted gadget in a category where Amazon was cheaper next week. PayPal processed payments for large merchants whose customers belonged to those merchants, not to PayPal. Its take rate fell from 2.89 percent in 2015 to 1.67 percent by late 2024, and management eventually chose to slow the line and improve the economics.

We are not claiming this pattern caused the outcomes on its own. Housing markets moved, retail competition intensified, and each company had its own circumstances. What we can say is that the pattern was present in every one, and that none of them managed to recover the economics somewhere else.

The more useful result is what we did not find. We went looking specifically for a company that combined very thin margins, one-time transactions, no strong repeat relationship and no durable cost or information advantage, and that still compounded over a long period. Across two deliberate searches covering roughly two dozen companies, we could not find a clear one.

That does not mean such a company cannot exist. It means we looked for it on purpose and came up empty.

Two cautions

Something carrying forward is not automatically enough

Peloton

The subscription was genuinely excellent, around 68 percent margin with low cancellation. Management was right that shifting the mix toward it improved every ratio in the business.

But hardware sales are what put new people into the subscription. As hardware slowed, subscribers peaked at 3.11 million and started drifting down. A valuable base still shrinks if nothing keeps filling it.

Charles Schwab

Schwab dropped trading commissions to zero in 2019 and made its money on the cash customers left in their accounts. For a while this was one of the best trades in financial services.

Then rates rose and customers could earn far more in a money market fund. Deposits fell 17 percent in 2022 and were down another fifth by the end of 2023. Schwab came through it, but the episode makes the point. What carries forward only helps if the customer cannot easily take it somewhere else.

Before you launch

Before chasing the volume, ask these questions

Will the same customer come back?

If not, who or what does. It may be a supplier, an agent, a lender or a merchant. In title insurance the buyer never returns and the business is fine, because the referrer does.
What exists after the transaction that did not exist before?

A customer, a balance, a membership, a supply agreement, a record, a route, a piece of distribution. Name the specific thing. If you cannot name it, that is the answer.
Does the next transaction get easier, cheaper or more profitable because this one happened?

This is the cleanest version of the test. A title search on a property gets cheaper. A house purchase does not.
Does this product need its own sales team, marketing budget or infrastructure?

If yes, a low gross margin is more expensive than it looks, because operating costs arrive behind it. If no, a 38 percent margin sold to an existing customer can beat an 80 percent margin you have to go out and win.
Can you measure the thing you say makes this worthwhile?

If the answer is engagement, lifetime value or strategic value, decide now which exact number proves it and when you expect to see it. The benefits that worked were measured. The ones that stayed vague were harder to see in the results.
Who controls the economics of what remains?

Costco sets its membership fee. Schwab does not set money market rates. If the customer can easily move the value somewhere else, it may not be as durable as it looks.

Early signs

What showed up before the outcome was obvious

Revenue rises faster than gross profit

A lot more activity producing very little extra value. Groupon showed this for years. The three big drug distributors added over $100 billion of revenue after 2015 while their combined gross profit dollars fell about 12 percent.

Gross margin falls and operating margin falls with it

The new product is carrying its own cost base as well as its own thin price. This was the more concerning pattern in the companies we studied.

Gross margin falls while operating profit improves

This one is healthy. It means the new product is riding a cost base someone already paid for. Shopify ran this pattern for three straight years.

The new product needs its own company inside the company

Warehouses, inventory, financing, field operations and a separate support function. Each of those makes a thin margin much harder to absorb.

The benefit is described but never counted

The benefits that worked were measured. Costco reports membership fees. Robinhood reports cash balances. The ones that stayed vague were harder to see in the results.

The buyer may not be the real customer

Sometimes the relationship that matters is with a merchant, a supplier, an agent or a lender who chooses on the buyer's behalf. Measuring the wrong side makes a healthy business look fragile, or the reverse.

In practice

If you are weighing one of these decisions

Write down what is supposed to carry forward. One sentence, before launch, naming the specific thing.
Decide how you will know. Pick the number and the date you expect to see it move.
Watch gross profit and operating profit, not revenue or volume. Revenue is the easiest thing for a low-margin product to produce.
Separate the two kinds of growth. Some volume builds something you can use again. Some has to be won from scratch every time. They should not sit in the same line on a dashboard.
Agree in advance what would make you stop. Several companies only repriced or exited after the mix had already shifted.

A low-margin product is not automatically a weak product. Several of the strongest companies in this study run products that make very little on each transaction, deliberately, and have done for years.

The difference is that the transaction leaves them with something useful for the next one.

The transaction is not the asset. What survives the transaction is.

Before chasing more volume, ask what will still be yours after the volume is gone.