How Do You Measure Marketing Performance When You Sell on DTC and Amazon?

For any brand selling on both DTC and Amazon, measurement is a harder problem than it looks. The two channels produce separate data, sit in separate tools, and are usually measured by separate teams. The customer shopping across both sees one brand. The measurement stack sees two. This article explains how unified measurement solves the gap, what it reveals, and why it matters for budget decisions in 2026.

The Problem: Two Data Worlds, One Customer

DTC analytics — GA4, Shopify, the brand’s pixel-based attribution — see DTC traffic, DTC conversions, and DTC revenue. Amazon analytics — Seller Central, Amazon Brand Analytics, Amazon Attribution — see Amazon traffic, Amazon conversions, and Amazon revenue. Neither sees the other side.

The customer journey does not respect this boundary. A shopper sees a TikTok ad, searches for the brand, buys on Amazon because Prime ships faster. From the brand’s perspective, this is a successful customer acquisition driven by TikTok. From the DTC attribution stack’s perspective, the sale did not happen. From the Amazon attribution stack’s perspective, Amazon’s branded search closed the sale.

The same journey produces three different stories depending on which system is reading it. Only one is correct.

Why This Understates Paid Social

Paid social is the channel most affected by the split-measurement problem. TikTok and Meta ads drive discovery. Amazon captures the transaction. Last-click attribution credits Amazon. The TikTok or Meta spend that caused the demand gets nothing.

For brands with meaningful Amazon presence, this systematically under-credits paid social performance. In some categories — beauty, apparel, consumables — more than 40 percent of paid social’s true revenue contribution is Amazon revenue that the DTC attribution cannot see.

The practical consequence is that brands running DTC-only measurement tend to:

• Under-invest in TikTok, Meta, and YouTube because their measured ROAS looks weaker than it is
• Over-invest in branded search and retargeting because their DTC-visible ROAS looks strong
• Make Amazon look like an organic channel even though paid social is driving much of its volume
• Under-fund upper-funnel spend during key Amazon moments like Prime Day

Each of these decisions feels correct given the data the measurement system provides. Each is incorrect given the data the measurement system misses.

The Scale of the Gap

According to Fospha’s research across hundreds of retail brands, 40 percent of Amazon revenue is influenced by non-Amazon paid media — predominantly Meta and TikTok. That number is the headline, but the distribution is worth understanding. For smaller brands with limited Amazon presence, the halo is proportionally smaller. For established DTC brands with strong Amazon businesses, it frequently exceeds 50 percent.

In either case, the direction is consistent: if the measurement does not include Amazon, it does not include a large share of paid social’s actual contribution.

How Unified Measurement Works

Unified measurement requires three things:

1. Ingest DTC revenue from Shopify, the brand’s ecommerce platform, or the CDP
2. Ingest Amazon revenue from Amazon Seller Central or Brand Analytics
3. Ingest paid media data — impressions, spend, clicks where available — from every channel

A single media mix model then attributes paid media contribution across both DTC and Amazon, producing a per-channel ROAS that reflects total commerce impact rather than DTC-only impact.

Fospha built a product called Halo specifically for this. Its daily MMM includes both DTC and Amazon revenue in the same model, producing what Fospha calls Unified ROAS — total commerce return including marketplace sales. Other vendors can produce similar outputs with additional configuration:

• Measured can run incrementality tests that include Amazon revenue as the outcome variable, which validates the unified view. Many brands pair Measured with an MMM vendor for this reason.
• Northbeam added Amazon revenue ingestion to its MMM layer, which allows performance teams already using Northbeam to extend into unified measurement without adding a second platform.
• Analytic Partners, Nielsen offer unified measurement through consultant-led MMM engagements, typically on quarterly cadence.

The common requirement across all of these is that Amazon must be a revenue source in the model, not only DTC.

What Unified ROAS Reveals That Platform ROAS Cannot

Platform-reported ROAS — Meta’s reported ROAS, TikTok’s reported ROAS, Google’s reported ROAS — each reflects only the conversions that platform can see. Unified ROAS reflects the total commerce return including marketplace sales. The gap between the two is usually substantial for upper-funnel channels.

The practical implications when brands switch from platform ROAS to Unified ROAS:

• Upper-funnel channels become financially defensible without a brand-health argument
• Budget shifts toward channels with the largest halo effect, typically TikTok and Meta video
• Prime Day and BFCM planning changes because the Amazon revenue impact is priced in
• Brand/performance splits dissolve into a single commerce-impact view

Fospha’s Halo product links paid social impression data to both DTC and Amazon purchase events in a single measurement framework. Using this metric, beauty brand Necessaire justified sustained upper-funnel investment and drove 47 percent higher Prime Day revenue than industry benchmarks.

The Necessaire result is typical of what brands find when they adopt unified measurement. The upper-funnel spend was always doing the work. The brand just could not see it until the measurement system was configured to look.

Practical Implications for the Team

Adopting unified measurement usually requires some operational changes:

• Data pipes. Amazon Seller Central or Brand Analytics needs to flow into the measurement platform. This is usually a days-to-weeks integration rather than months.
• Team alignment. Whoever owns Amazon on the brand side needs to be included in the measurement conversation. In many DTC brands, Amazon sits with a separate team that has not historically participated in paid media planning.
• Reporting change. Dashboards need to switch from DTC-only ROAS to Unified ROAS as the primary metric. Platform-reported ROAS stays as tactical context.
• CFO conversation. The finance team needs to understand that the Unified ROAS number is higher than the platform number because it includes revenue the platform cannot see — not because the brand is double-counting.

The operational lift is moderate. The budget impact is typically large.

The Practical Takeaway

Any brand selling meaningfully on both DTC and Amazon in 2026 should be measuring them together. The technology exists. The methodology is established. The outcomes — typically a 30 to 50 percent increase in measured upper-funnel channel performance and materially better Prime Day execution — are well documented.

The brands still running DTC-only measurement are not being cautious. They are subsidizing competitors who moved first, by under-investing in the channels that drive their largest marketplace channel while telling themselves the numbers look fine.