Marketplace vs. own online store: a study of EU businesses

“Marketplace vs. ecommerce” is a common way to frame this decision, but a marketplace is itself a form of ecommerce. The useful comparison is between taking orders through a business’s own website or app and taking them through an external marketplace. A business can use either route, or both.^[2]

This study follows those ordering routes. An own store means the business receives web orders through its own website or app. Marketplace selling means it receives them through a marketplace website or app. The distinction is about where orders are placed, not whether the business coded its checkout software itself. It also applies to services and business customers, not just retail products.

A firm classified as selling only through marketplaces may still have an informational website. It may also take orders offline or through other electronic systems. “Marketplace-only” here means no web sales through its own website or app. It does not mean that every sale the business makes passes through a marketplace.^[1][2]

Keeping those definitions straight prevents a practical mistake. Building a website, opening an online store and changing the mix of orders are different developments. A team should check which of them it actually needs before treating a new website as the answer.

Among all eligible small EU firms, 8.95% sold through a marketplace in 2024. For large firms, the figure was 14.16%. That is a measure of adoption: a larger proportion of large firms used this route. It includes firms with no web sales at all in the denominator.^[1]

Now restrict the question to firms that actually sold through a marketplace. Among small firms in that group, an estimated 83.8% received at least 20% of web sales there. Among large marketplace sellers, an estimated 65.8% did. This comparison asks how often marketplaces contribute at least a fifth of web sales among the businesses using them, rather than how widely the route is adopted.

The small-firm calculation is 7.50 divided by 8.95, multiplied by 100. Eurostat reports that 7.50% of all eligible small firms both used marketplaces and received at least a fifth of web sales through them. Dividing by the 8.95% that used marketplaces gives the 83.8% result. Medium-sized firms sit between the two groups, at an estimated 77.5%.^[1]

The threshold is important. A firm receiving 20% of its web sales through marketplaces qualifies, just as a firm receiving 80% does. The statistic does not tell us that most of an individual firm’s sales come from marketplaces. Nor does it say marketplaces collect 83.8% of the small-business market’s sales value. We are counting businesses that meet a condition.

For an owner, adoption figures can provide context about common routes to customers. They cannot substitute for the business’s own sales split. A channel used by relatively few businesses could still account for a meaningful portion of orders in the particular business you manage.

One possible explanation is that small marketplace sellers are more likely to have no own-store sales. In the EU figures, 33.7% of small marketplace sellers used marketplaces as their only web-sales route, compared with 18.2% of large marketplace sellers. These firms necessarily receive all their web sales through marketplaces, so they all meet the 20% threshold.^[1][6]

We therefore repeated the calculation only for firms selling through both routes. After removing marketplace-only sellers from the threshold group, an estimated 75.5% of small firms using both routes still received at least a fifth of web sales through marketplaces. The corresponding estimates were 69.5% for medium-sized firms and 58.2% for large firms.

For small firms, the calculation is (7.50 − 3.02) divided by 5.93, multiplied by 100. The first two numbers are the percentages of all small firms meeting the threshold and selling only through marketplaces. The final number is the percentage selling through both their own website or app and a marketplace. Each input uses the same business population.^[1]

This check narrows the interpretation. The small-versus-large difference is not explained solely by the proportion using marketplaces as their only web-sales route. It remains among businesses with own-store sales too. That does not show what caused the difference: customer demand, industry, product mix and other circumstances can vary between the groups.

An own store therefore should not be treated as evidence that marketplaces have become marginal. Equally, a substantial marketplace contribution is not evidence that an own store has failed. The two routes may be serving different purchases. The data tell us to examine that possibility, not which route deserves the next euro of spending.

The small-firm share reaching the 20% threshold was higher than the large-firm share in 23 of the 26 EU countries with both estimates available. Malta, the Netherlands and Slovakia were the exceptions. Germany’s large-firm threshold estimate is unavailable and flagged for low reliability; it is missing, not zero.^[1]

Restricting the comparison to firms with both sales routes also gives a small-firm lead in 23 of 26 countries, but the exceptions change to Bulgaria, Malta and the Netherlands. The distinction is a reason to show both comparisons. It would be misleading to assume that one description fits every market.

Country adoption and sales-share figures can diverge too. In Lithuania, 34.74% of small firms sold through marketplaces, compared with 9.22% in Ireland. Yet among their marketplace sellers, the estimated share reaching the threshold was 82.4% in Lithuania and 91.4% in Ireland. More widespread adoption does not necessarily mean a larger share of marketplace sellers cross the 20% web-sales threshold.

These are descriptive comparisons, not a ranking of marketing quality. The countries have different mixes of industries, and this analysis does not adjust for them. The Netherlands carries source flags for a break in series on relevant small-firm measures; we retain those flags and use the figures as a single-year comparison. The country table preserves unavailable values and the source notes.

Consider a fictional ceramics workshop with 12 people employed. During one year it receives €420,000 in orders through its own store and €180,000 through marketplaces. Its web sales total €600,000. It also makes €400,000 in offline sales. These amounts are invented to explain the calculation, not drawn from a customer or presented as expected results.

Marketplaces account for 30% of its web sales: €180,000 divided by €600,000. The business meets the study’s 20% threshold even though its own store accounts for the other 70%. Relative to its €1,000,000 in total sales, the marketplace share is 18%. Both descriptions are correct because they use different denominators.

Suppose the team is preparing posts for a new range. The fact that it has an own store does not settle where those posts should lead. It could review the product details, ordering steps and customer questions at each destination. A post promising help choosing a size should lead somewhere that supplies that help; a post advertising a particular offer should lead to the place where that offer can actually be ordered.

The 30% share alone cannot answer whether the marketplace is worthwhile. The team would also need its costs, margins, returns and the staff time associated with each route. Those details are not in this study. The worked example is useful because it separates three questions the team can otherwise blur together: where it sells, how much each route contributes, and whether the economics work.

Start with the customer’s immediate question. Someone comparing specifications may need a detailed product page. Someone responding to a time-limited offer needs the exact price, availability and ordering conditions. Someone arranging a service may need booking details. Review the actual pages you can maintain before deciding which destination best answers that question.

Compare both destinations against the same practical checks. Can a person find the product or service promised in the post? Are the price, availability and delivery or booking details current? Can they complete the next step on their phone? A marketplace might be the better fit for one campaign and an own store for another. The choice should follow what the customer needs to do.

For the fictional ceramics workshop, a post explaining how to care for a glaze could lead to an own-store page that combines care advice and the relevant products. A post promoting a bundle offered only through a marketplace should lead to that offer. In each case, the destination follows the promise in the post. There is no reason to send every customer through both routes.

First, use existing order records to list web-sales value by ordering route for the same period. Keep own-store orders and marketplace orders separate, then calculate each as a share of their combined web-sales total. Keep offline sales in a separate line so that a share of web sales is not mistaken for a share of all business sales.

Use a consistent sales-value basis. A marketplace payout after fees is not directly comparable with an own-store order total before fees. Mixing the two can make one channel look smaller for an accounting reason. Reconcile the sales totals first, then compare the relevant fees, returns and other costs separately.

The fictional workshop’s review would start with €420,000 and 70% of web sales for its own store, and €180,000 and 30% for marketplaces. The next entries would be its actual costs and recurring customer questions for each route. Check the records for any missing cost before assigning it a value. This creates a more useful decision than treating the study’s 20% cutoff as a target every business should meet.

Finally, choose one concrete action for the next campaign: correct a missing delivery detail, improve the product explanation or change the destination to match the offer. Use the business’s existing measurement to review the result. The order route is not the same as the discovery source: someone may see a social post and order elsewhere later. A change in channel sales alone does not prove that the campaign caused it.

We analyzed Eurostat’s isoc_ec_esels dataset, using the 2025 survey observations on sales in 2024. The population covers enterprises with at least 10 employees and self-employed persons in manufacturing, construction and the other covered non-financial industries. It excludes firms with fewer than 10 people employed and is not a dedicated survey of online retailers or creators. We use Eurostat’s supplied EU aggregate, not an average of country percentages.^[1][2][3]

The main measure divides the percentage of enterprises receiving at least 20% of web sales through marketplaces by the percentage with any marketplace web sales. For firms using both routes, we subtract marketplace-only sellers from the threshold numerator and divide by sellers using both. The inputs use matching country, year, industry scope, business size and percentage-of-enterprises units. Source values, formulas and quality flags are included in the downloads.^[1]

Germany’s published large-firm threshold estimate is unavailable because it is flagged for low reliability. Eurostat includes low-reliability national estimates in its EU aggregates, so we retain the published EU figures while leaving Germany out of the country counts that require this estimate. The country table shows an unavailable value, not zero. This differs from Eurostat’s treatment of confidential national values, which it excludes from EU aggregates.^[3][6]

We checked the calculations against Eurostat’s alternative percentages of web sellers and checked whether the channel categories reconciled. One source discrepancy remains: for Lithuanian medium-sized firms, combining own-store and marketplace categories yields 43.81% with web sales, whereas the published web-sales total is 43.85%. A separate retrieval reproduced it. We preserve the values and do not present that composition as an exact total. The discrepancy does not enter the threshold-to-marketplace calculation or the small-versus-large country comparison.^[1]

The published inputs are rounded to two decimal places. We tested each headline calculation against values within 0.005 percentage points of each input. The small-versus-large headline differences and reported country directions survive this rounding check. These ranges concern display precision, not survey confidence intervals. The extract does not supply the respondent data or survey covariance needed to estimate uncertainty for our derived comparisons.

This is a snapshot of reported sales routes for one year. It cannot establish why firms choose them, whether one route is more profitable, or what would happen if a firm changed its mix. Earlier Eurostat and OECD work has already compared ecommerce channels. Our contribution is the calculation among marketplace sellers, the further comparison among sellers using both routes, and the country checks on the latest sales year available in this extract.^[2][4]