Ecom Heads Newsletter Preview
Ecom Heads
A marketplace can find your next customer. Measure what happens after the first order.

A July survey of 1,000 U.S. consumers found that 44% first buy from a new brand on a marketplace such as Amazon or Walmart.

Take the 44% with some caution. Radial commissioned the Dynata survey and sells fulfillment services. The result is self-reported consumer research. The release did not include the full questionnaire or weighting.

Your own channel data can tell you whether that behavior is showing up in your business.

If marketplace revenue sits in a separate channel report, you may be judging a first customer order as another sales total. That leaves the acquisition question unanswered: did the marketplace find a customer you would have missed, and what was that customer worth after fees, fulfillment, returns, support, and repeat behavior?

John Sciacchitano
🤝 Connect with John on LinkedIn
Say hey, I'll respond!
Connect →

The 44% finding

Dynata collected the survey in March 2026 from 1,000 U.S. adults. Forty-four percent said they first buy from a new brand on marketplaces. Ten percent selected online retail sites such as Target or Macy's, and 3% selected social shopping channels such as TikTok Shop.

The same research found that 38% were shopping on marketplaces more often than the year before. Use that finding as a reason to inspect your acquisition mix. Keep it out of forecasts until your own data supports it.

FIRST PURCHASE FROM A NEW BRAND
44% via marketplaces
Radial-commissioned Dynata survey of 1,000 U.S. adults, collected in March 2026.

Amazon is measuring past the first order

Amazon's advertising reports provide a useful example. Amazon defines a new-to-brand purchase as a first purchase from that brand within a 12-month lookback window. The metric is available for Sponsored Brands and Display, though Sponsored Products does not offer the same new-to-brand reporting.

Amazon also has long-term sales and long-term ROAS metrics. They estimate 12-month sales value from first-time detail-page views, branded searches, add-to-carts, and purchases using historical behavior on Amazon.

These metrics stay inside Amazon. A shopper can qualify as new to the brand on Amazon after buying from the same brand elsewhere. Still, the measurement direction is useful. The first order is only the opening entry in the cohort.

Define marketplace-first before you count it

Every source uses a different boundary.

Amazon's new-to-brand window looks back 12 months on Amazon. Your DTC system may use a customer's first recorded order across the full account history. A retailer partner may provide only aggregate sales.

Put the definition beside the number in your report. Record the platform, lookback window, ad products included, country, and missing channels. That note will keep a change in reporting scope from looking like a change in customer quality.

Customer-level matching across a marketplace and your direct store may be unavailable or restricted. Use the data the channel permits. Aggregate cohorts, platform-provided new-to-brand metrics, controlled launches, branded search, direct traffic, and post-purchase survey responses can still show direction.


Build the contribution view

Start with net sales after refunds. Then subtract:

  • Cost of goods
  • Marketplace and payment fees
  • Marketplace advertising
  • Fulfillment, storage, inbound freight, and returns processing
  • Concessions and a consistent support-cost estimate

The remainder is cohort contribution. Divide it by new-to-brand orders when the platform supplies that count.

This makes two periods comparable even when gross marketplace sales rose. A channel can add orders while contribution per acquired customer falls because ad spend, storage, returns, or support moved in the opposite direction.

Keep the DTC comparison honest. Use the same cost categories, the same time window, and the same refund treatment. DTC has its own payment, fulfillment, support, and acquisition costs.

Follow the customer past day one

Measure repeat orders, sales, and contribution at 30, 60, and 90 days when the category supports those windows. A replenishable product may show useful repeat behavior quickly. A durable product may need a six- or 12-month view.

Amazon's long-term sales metric can add a modeled platform view for eligible advertisers. Keep it separate from realized cohort contribution. A model and a completed order answer different questions.

Owned-channel movement usually needs aggregate evidence. Watch branded search, direct traffic, permitted email signups, and post-purchase survey responses around marketplace launches or growth periods. If volume allows it, compare regions, products, or periods with different marketplace exposure.

Give the pattern a few reporting periods before you trust it.

Check the experience behind the cohort

The Radial survey also found gaps in delivery, pricing, and inventory across marketplaces. Respondents reported slower or less reliable delivery on TikTok Shop at 16%, compared with 7% on Amazon and Walmart Marketplace. Different prices were most often reported on Amazon at 17%, while stockouts were most often reported on Walmart Marketplace at 16%.

Those figures are survey responses, and every category will behave differently. They point to useful checks:

  • Is the same SKU described consistently?
  • Does marketplace pricing create a surprise when the shopper finds your site?
  • Are inventory and delivery promises accurate?
  • Can support resolve a marketplace order without sending the customer in circles?

Radial's results suggest that shoppers still hold the brand responsible when the order goes wrong.


Build this report this week
1
Choose one completed cohort. Start with a month or quarter that has enough time for returns and repeat orders to appear.
2
Write down each new-to-brand definition. Include the platform, lookback window, eligible ad products, and known gaps.
3
Calculate contribution by channel. Use the same cost categories for marketplace and DTC orders.
4
Add repeat checkpoints. Measure 30-, 60-, and 90-day behavior where the product cycle makes sense.
5
Add aggregate brand-demand signals. Track branded search, direct traffic, permitted signups, and survey answers around changes in marketplace exposure.
6
Audit the customer experience. Compare SKU content, price, inventory, delivery, returns, and support across channels.

Run the report for several cohorts before making a large channel decision. One launch month can be distorted by discounts, Prime Day, inventory gaps, or a single hero SKU.


When the data stays messy

Low order volume can make repeat rates swing. Long repurchase cycles delay the answer. Marketplace definitions may miss purchase history from every other channel.

Use wider time windows and fewer segments when the sample is thin. A controlled region, SKU, or launch-period comparison can help when customer-level matching is impossible. Keep causal claims narrow.

Some marketplaces will be profitable demand capture with little visible migration to an owned channel. That can still be a good business. Judge that outcome on contribution and strategic control. Email-list migration can remain a separate secondary metric.

The bigger point

Marketplace revenue deserves the same customer-acquisition analysis you apply to paid media.

The Radial survey puts 44% of first purchases from new brands on marketplaces. Your category may be far above or below that. Pull your new-to-brand data, calculate the contribution, and follow the cohort long enough to learn what the first order produced.

Gross revenue tells you what the channel sold. The cohort tells you whether the customer was worth acquiring.

Sources

John Sciacchitano
🤝 Connect with John on LinkedIn
Say hey, I'll respond!
Connect with John →
Talk soon,
John Sciacchitano
Ecom Heads: Scale or Die Trying

Keep Reading