Skip to content
Novandri.

E-commerce vs Brick and Mortar: The Economics of Retail Channels

Online and physical retail have different cost structures, not different laws of economics. Read the 2026 Census data and a margin worked example.

Rachmadwipa Novandri8 min read

Retail channels are often discussed as if one were destined to replace the other. The more useful comparison is structural: each channel carries a different set of costs and offers a different set of benefits to the customer. A business that understands those differences can choose a channel mix deliberately rather than following a trend.

Direct answer:

E-commerce and physical retail differ in rent, labour, inventory, fulfilment, and the way customers discover and evaluate products.

The U.S. Census Bureau reported that e-commerce accounted for 17.1% of total retail sales in the second quarter of 2026, with e-commerce growing faster than total retail year over year. The boundary: a channel share is an aggregate across all retailers, not a forecast for a specific category, and the two channels are increasingly combined rather than chosen exclusively.

17.1%E-commerce share of total retail sales, Q2 2026U.S. Census Bureau
340.2 billionSeasonally adjusted U.S. retail e-commerce sales, Q2 2026U.S. Census Bureau
12.2%Year-over-year growth in e-commerce salesU.S. Census Bureau
A comparison: online retail's fulfilment costs versus physical retail's rent and labour.
Conceptual: the channel mix should be deliberate. A channel decision made on revenue alone ignores them.

What the official data says

The Census Bureau's Quarterly Retail E-Commerce Sales release for the second quarter of 2026 reports:

  • U.S. retail e-commerce sales of 340.2 billion, seasonally adjusted but not price adjusted, up 3.8% (±0.4%) from the first quarter.
  • Total retail sales of 1,986.5 billion, up 2.9% from the first quarter.
  • Year over year, e-commerce rose 12.2% while total retail rose 6.7%.
  • E-commerce accounted for 17.1% of total sales.

Two readings matter. First, e-commerce is growing roughly twice as fast as total retail, which is why its share rises. Second, it remains under a fifth of the total — the other four fifths are physical or non-store channels that are not captured by the e-commerce definition. "Online is the future" is a directional claim that conceals the present structure.

The release also notes that after 28 September 2026 the report no longer contains the most up-to-date estimates, with revisions due in the third-quarter release on 19 November 2026. As of this writing, that revision has not been published, so the figures above are the current published estimates with a stated revision pending. Checking the release date of a statistic is part of using it honestly.

The cost structures differ

Physical retail carries rent, in-store labour, utilities, and the cost of displaying inventory. In exchange, the customer can inspect the product, take it immediately, and receive human assistance. The store is both a fulfilment point and a marketing channel.

E-commerce avoids store rent and sales floor labour but adds fulfilment, packaging, last-mile delivery, digital acquisition, and returns processing. The website is a storefront that never closes, but it must buy attention and solve the trust problem of buying something unseen.

Three economic consequences follow:

  1. Fixed versus variable cost. A store commits to rent before knowing demand; an online operation can scale fulfilment more flexibly but pays per order. The break-even point differs.
  2. Returns asymmetry. Clothing bought online has high return rates because fit cannot be verified. Returns reverse a completed sale and add a handling cost that a store visit avoids.
  3. Discovery economics. A store benefits from foot traffic; a website must pay for search, ads, or content. The acquisition cost is more measurable online and not necessarily lower.

NRF press releases are a useful primary source for what the retail industry itself reports, but they are claims by an industry body. Treat them as a perspective on conditions, not independent measurement; the Census figures are the better anchor for magnitudes.

A worked example of channel margin

The following numbers are hypothetical and illustrative.

A product sells for 100 and costs 45 to manufacture.

In-store: the retailer pays 15 in rent and staffing per unit sold, loses 3 to shrinkage, and accepts no return cost. Contribution is 100 − 45 − 15 − 3 = 37.

Online: the retailer pays 12 in fulfilment and shipping, 8 in digital acquisition, and 5 in returns and reverse logistics (assuming a 20% return rate with a net cost per return). Contribution is 100 − 45 − 12 − 8 − 5 = 30.

On these assumptions the store is more profitable per unit. But the comparison changes with volume and product type:

  • If online acquisition falls to 4 through repeat purchases, online contribution rises to 34, nearly matching the store.
  • If the product is bulky and store space is constrained, the store's 15 rent allocation understates the opportunity cost of the shelf.
  • If the customer buys online and returns to a store, the two channels share a cost that a pure-channel model misallocates.

The lesson is not that one channel wins. It is that the margin depends on variables — acquisition, returns, rent per unit — that differ by product and by customer behaviour.

Damodaran's valuation discipline is useful here: the relevant question is whether the channel produces adequate returns on the capital it ties up, not whether its revenue grows faster.

What the customer is actually buying

A channel comparison that counts only costs misses the fact that the two channels deliver different bundles.

Immediate possession. A store customer leaves with the product. An online customer accepts a delay in exchange for selection or convenience. The value of that delay varies by category — urgently needed items favour the store, while a broad catalogue favours online.

Inspection and certainty. For products where fit, texture, or appearance matters, a store reduces the risk of a wrong choice. That is why high-return categories online are precisely the ones where inspection matters most.

Assistance. A knowledgeable salesperson can resolve uncertainty that a product page cannot. This is a real service with a real cost, and it is why complex purchases often still favour physical channels.

Effort. Searching online can be faster than driving to a store, or slower if the catalogue is overwhelming. The direction depends on how well the customer knows what they want.

These differences explain why the same retailer can rationally maintain both channels. They are not competing versions of one offering; they are different bundles of cost and benefit, and customers choose between them by category and circumstance.

Where the binary breaks down

Three developments make the either/or framing misleading:

Buy online, pick up in store uses the store as a fulfilment node and reduces last-mile cost. Ship from store uses existing inventory instead of a separate warehouse. Both blur the cost structures above.

Marketplaces introduce a third party that owns discovery and sometimes fulfilment, changing the acquisition and delivery costs for the seller.

Channel-specific pricing means the same product may carry different prices online and offline, which changes customer behaviour and makes aggregate comparisons harder.

A decision test

  1. Define the category and channel before using an industry statistic; the 17.1% figure is an all-retail aggregate.
  2. Build a contribution margin for each channel with explicit rent, labour, acquisition, fulfilment, and return costs.
  3. Check the return rate by product type before committing to online.
  4. Include the customer's perspective — inspection, immediacy, and assistance have value that does not appear in the cost column.
  5. Model mixed channels where the store can serve as a fulfilment point.
  6. Check the release date of any market statistic and note pending revisions.
  7. Judge the channel by return on capital, not by revenue growth alone.

For the operational model, see Omnichannel Retail; for the valuation frame, see Corporate Finance Essentials.

A counterargument

A reasonable critic can object that the article's structural comparison rests on hypothetical costs while its only quantitative source is an aggregate. The Census release supplies a share and a growth rate, not channel-level profitability, and the article correctly warns that the 17.1% figure cannot be applied to a specific category. NRF is an industry body whose press releases the article itself treats as a perspective rather than independent measurement, and Damodaran's discipline concerns returns on capital that the cited sources do not provide at channel level.

The objection is not that the two channels differ in cost structure; it is that a decision test built on return on capital cannot be executed with the evidence on hand.

Under conditions where store and online costs are shared, the framework distinguishes the channels conceptually but cannot quantify the comparison it recommends.

References & Further Reading