Background
Traditional multi-touch attribution only credits clicksâignoring the powerful impact of ads users see but donât click. On Safari, iOS, or privacy-first browsers, up to 60% of your brand ad exposure goes unmeasured. That means your display, video, and social awareness campaigns appear âineffectiveââeven when theyâre driving real sales.
This model is particularly suitable for DTC brands that invest budgets in short-video ads such as TikTok ads, Youtube, and Meta ads. Because these ads cannot track conversions through clicks at all.
Introducing Synthetic Impressions by Attribuly
Our breakthrough solution restores view-through attribution in a cookieless world. Using advanced machine learning and your actual site behavior data, Attribuly intelligently estimates the probability that a visitor saw your adâeven without tracking IDs.
How it works:
We analyze your aggregated ad delivery data (impressions by campaign, geo, device).
We observe orders where the first touchpoint comes from Direct, Email, Organic search, or Paid search with brand keywords.
Our model identifies behavioral patterns linked to ad exposureâand synthesizes probabilistic impressions for every visitor.
To view the Full Impact model,
Go to all attribution
Click "Full Impact Model"
You may notice a decrease in contributions from Direct, Email, Organic search, and Search ads. You can compare the data differences between different models.
Currently Only Meta & Google ads are supported, we are rolling in more ad networks.
Introducing dynamic attribution model with machine learning
Traditional models are mostly rule-driven, treating cross-channel touchpoints as isolated events, making it difficult to capture the true impact relationships between channels. Attribuly's machine learning (ML) model depicts cross-channel relationships through advanced algorithms; it is more concerned with whether adding or reducing a certain channel in the entire user path will have a significant impact on the conversion rate. Obviously, touchpoints that have a significant impact should receive more credit.
This new algorithm completely abandons the fixed way of giving credit based on the position of touchpoints and can more truly reflect the contribution of a certain touchpoint. For example, brand keyword ads have a very high contribution from the Last click perspective. However, under the Full Impact model, the contribution of brand keyword ads will be significantly reduced because even without this ad, it does not affect the user's purchasing intention. The following figure describes how the Full Impact model focuses on the incremental contribution of a certain ad.
You may be concerned about the differences between Full Impact and Google Analytics 4's DDA (Data Driven Attribution). Here are some of our summaries.
Please note that Full Impact is currently in the Beta phase and may has wrong result. Merchants on the Enterprise plan and those with a monthly GMV exceeding 100K USD are welcome to contact us for a free trial.
FAQ:
Q: Attribuly states that Full Impact measures the contribution of ad impressions to conversions and can capture ad browsing behavior. Can we confirm that these impression data are mainly obtained through direct integration with ad platforms (to get platform-side impression data)?
A: Yes, we directly retrieve various ad platform data via API for model training, including information such as impressions, interactions, and geolocation.
Q: For which scenarios is the modeling of non-clicked impression impact applicable?
A: It mainly applies to orders where the first touchpoint in the user's behavior path is Direct, organic search, brand term search ads, Email, and other channels with obvious touchpoint information loss.
Q: For KOL/content-type exposures on platforms like YouTube: Can it be included in Full Impact or attribution analysis? I
A: If the YouTube KOL video placement is an ad, its data can be included in Full Impact. If it is just organic YouTube content views, it cannot be included in the Full Impact prediction scope.