Sellforte launches pixel for experiment-calibrated attribution

Add up the sales every ad platform claims for a given week and the total can come out higher than the number of orders the store actually took. Sellforte Incremental Pixel is the company's own website tracking layer, and it counts every order once. Sellforte then calibrates attribution with causal evidence from incrementality experiments and marketing mix modeling (MMM).

Sellforte announced the launch in a release distributed through PR Newswire. The pixel is available now inside Sellforte Incremental Attribution, and the company aimed it at ecommerce brands in the mid-market and direct-to-consumer segments. For the performance marketing lead at one of those brands, the question is simple. Can daily bid and budget moves run on numbers that have been checked against experiments, instead of on what each platform reports about itself?

Juha Nuutinen, CEO of Sellforte, said ad platforms are "very good at taking credit," which is why their combined figures often total more sales than a brand actually made. Performance teams, he said, need daily figures they can rely on, and the product lets them optimize toward the results marketing caused rather than toward whatever happened closest to the purchase.

Two errors, two separate fixes

Double counting happens when several platforms each claim the same purchase. The pixel handles that by anchoring attribution to orders that actually happened, so each conversion enters the measurement once and its credit gets split across the touchpoints Sellforte can observe.

The second problem is harder. Attribution tends to reward the channels closest to the purchase, whether or not they caused it. Deduplication doesn't fix that, because one correctly counted order can still go to the wrong channel. This is where the experiments come in. Results from Conversion Lift and GeoLift tests reweight the credit each channel receives. When a test shows the sales would have come in regardless, that channel's share goes down. The adjustment can also run the other way: Sellforte says prospecting and video get more credit when experiments show they drive more sales than browser tracking picks up.

MMM still guides the larger strategic budget split. Platform-reported results sit next to the incremental figures, so a team can see where each platform overstates or understates its contribution.

That split explains why the product exists. Experiments and MMM have usually been the slow, strategic tools: useful for setting a quarterly channel mix, too coarse to tell a buyer which ad set to pause on a Tuesday. Sellforte's pitch is that their findings can be carried down into the daily, ad-level view that performance teams already work in.

Brands can keep their GA4 or Snowplow setup

The detail that matters most for an evaluation is easy to miss. Brands don't have to install the pixel at all. According to the release, companies already running Google Analytics 4, Snowplow or a specialist attribution provider can send that data to Sellforte for calibration with incrementality experiments and MMM. Sellforte's product page puts it plainly: where the raw attribution comes from matters less than calibrating it toward incremental impact.

The pixel is optional; the product is the calibration step that marks a channel's credit down when a lift test shows the sale would have happened anyway.

For a brand that has spent years configuring GA4 events or building a Snowplow pipeline, that changes how big the decision is. It's a question of whether the calibration earns its place on top of data the team already trusts.

Brands that adopt the native pixel can get started with ready-made Google Tag Manager snippets and built-in tools for generating and validating tracking events. By default, the pixel does not collect user IDs, email addresses or payment data. Privacy and legal reviewers will want to see those defaults spelled out before any new tag goes live on a checkout flow.

What tracking can't see

Sellforte is upfront about the limits of any pixel. Its product page lists the gaps: consent choices, multiple devices, browser restrictions and ad views that never get a click all break up individual-level tracking. The company says it doesn't pretend to see those touchpoints. It uses experiments and MMM to estimate their effect in aggregate.

That is also where the approach has a weak point. Calibration can only be as current and as broad as the experiments behind it. A brand that runs lift tests on a few channels, or ran them months ago, could end up with daily figures that rest more on MMM, the method Sellforte itself positions for strategic allocation rather than ad-level moves.

Where the budget conversation starts

Sellforte says its platform serves the world's biggest retail and ecommerce brands, naming Grupo Soma, bonprix and Douglas. Incremental Pixel, by contrast, is built for mid-market and direct-to-consumer ecommerce brands. Those teams often run lean, with one or two people managing spend across several platforms that each report their own results.

For a team that has been allocating on platform-reported ROAS, the side-by-side view is the practical starting point. The columns where platform-reported and incremental figures split furthest apart show where the budget arguments will be, and which channels a lift test should cover next.