Data Stack for a Multi-Channel DTC Brand
The same customer and the same product now live in several systems, each with its own fees, dates, and SKU codes. This is the stage where a warehouse usually starts to earn its cost.
The short answer
Sound familiar?
- Channel totals do not add up to the bank.
- Amazon profit moves after the sale as fees settle.
- One product has a different SKU code on each channel.
- Someone reconciles the channels by hand every month.
Why it happens
- Fees arrive after the sale
- Marketplace fees can post days after the order, so recent days show the sale but only part of the cost. Early margin looks better than it will end up.
- One purchase creates several records
- A TikTok Shop order can be copied into Shopify and shipped by Amazon fulfillment. Those three records describe one $60 sale, and adding the purchase and its Shopify copy would show $120.
- Fees hide in separate fields
- In one TikTok Shop order traced on the blog, the platform commission field was 0.00 while affiliate, referral, and promotion fees still totaled 8.55 on a 25.50 sale. A report that reads only the commission field makes the channel look cheaper than it is.
How we fix it
- 1
Choose ingestion by fee coverage
Pick Daton or Fivetran by whether the connector loads the Amazon settlement and fee reports you need, not only orders. Land the data in BigQuery or Snowflake, whichever cloud you already use.
- 2
Model orders, refunds, and fees per channel
Build order, refund, and fee models for each channel, a SKU map with pack quantities and dated costs, and an order crosswalk that links copies and fulfillment records to one business order.
- 3
Write the rules down, then report
Record which date a refund uses and how estimated and settled fees are labeled. Report in Sigma if finance works in spreadsheets, Power BI in a Microsoft shop, or Looker Studio for a free scorecard.
What you end up with
- Order, refund, and fee models for each channel
- A SKU map with pack quantities and dated costs
- An order crosswalk so no purchase counts twice
- Daily KPIs and contribution margin by channel
2 weeks to minutes
to answer a cross-brand question for a multi-brand ecommerce portfolio
Read the case study
Common questions
When does a multi-channel brand need a data warehouse?
Usually when channel totals no longer add up to the bank and someone reconciles them by hand each month. At that point the fees, dates, and SKU codes differ enough across Shopify, Amazon, and TikTok Shop that one model is cheaper than monthly spreadsheets.
Can a packaged tool handle Amazon and TikTok Shop?
Sometimes. Polar or Peel can fit if Shopify still leads and their definitions match yours, and Peel reads Shopify and Amazon. Sellerboard fits if Amazon is most of revenue. As of November 2026, Triple Whale's TikTok Shop integration is US-only and leaves out platform and affiliate fees. Check current plans with each vendor.
Why is our combined revenue higher than the channel reports?
Often because one purchase reaches the warehouse as several records, such as a TikTok Shop order, its Shopify copy, and an Amazon fulfillment record. Keep sales channel, fulfillment provider, and source system as separate fields, and link each record to one business order before adding up revenue.
Which connector should we use for Amazon?
Check the exact objects each connector loads. You need settlement and fee reports, not only orders. Daton targets marketplace sources at low cost, and Fivetran covers a broad catalog with nothing to host.
Should we drop the packaged tool marketing uses?
Not necessarily. Many brands keep an attribution or retention tool for daily marketing views and load its data into the warehouse, which handles finance, margin, and anything that has to tie out.
Guides
Bring the numbers that don't add up
We'll trace them to the source with you and say where we'd start. No pitch deck.