Unified cross-channel dataset linking mobile, CTV, and CPG for closed-loop attribution.
Since 2018
Privacy-safe enterprise data
VOSB & SDVOSB
U.S. Veteran-Owned, Service-Disabled
SCC Transfers
EU/UK/CH per privacy policy
GDPR · CCPA · CPRA
Consent-first data sourcing
Overview
A unified dataset that deterministically links Mobile Ad IDs, Smart TV ACR exposures, and CPG purchase events into a single consumer graph. The Euclidean Feed makes true closed-loop attribution possible, connecting media exposure to real-world purchases across digital, TV, and in-store channels.
Want the buying guide?
If you’re evaluating vendors or writing an RFP, this guide lays out what to ask for (coverage, refresh, governance) and links to the best references.
A representative slice of the Euclidean Feed schema. The data dictionary shared during evaluation lists every field with its type, nullability, join keys, and refresh cadence.
Field
What it contains
source_record_id
Pseudonymous identifier for the source identity record.
candidate_record_id
Pseudonymous identifier for the candidate match record.
distance_score
Pre-computed similarity distance between the source and candidate records.
cluster_id
Identifier for the resolved group of related records.
How the Euclidean Feed Links Signals
The Euclidean Feed is GSDSI's flagship cross-channel identity product, built by deterministically linking three core data assets: Mobile Advertising IDs (MAIDs) from our device panel, Automatic Content Recognition (ACR) viewership events from smart TVs, and transaction-level CPG purchase records. The linkage methodology uses shared household graphs. Devices and TVs that consistently co-locate at the same residential address are grouped into household units, and purchase data is associated through retailer loyalty program matching and probabilistic address linkage. The end result is a unified consumer graph where a single household's CTV ad exposure, mobile app activity, web browsing, and grocery purchases are all connected, making true multi-touch attribution possible across the full marketing funnel.
Common Applications for Unified Attribution
CPG brands use the Euclidean Feed to prove that CTV advertising drives incremental product purchases at specific retailers, closing the loop from impression to checkout in a way that was previously impossible without single-source panel data. Media agencies use the cross-channel linkage to build unified reach and frequency reports, deduplicating audiences across CTV, mobile, and web to eliminate overexposure and optimize media plans. Retail advertisers connect ad exposure to store visits by linking MAID-based campaign impressions to GPS-derived foot traffic events at their locations. Financial analysts use the Euclidean Feed's multi-signal consumer profiles to build stronger alternative data models that combine viewership, engagement, and purchase behavior into a single composite indicator per ticker.
How buyers diligence cross-channel attribution graphs
Euclidean buyers should test whether the graph can link exposure, mobility, and purchase signals without double counting households or overstating deterministic identity. The pilot should be scoped around one campaign or cohort.
Confirm the identity logic behind MAID, CTV, household, and CPG joins.
Validate dedupe, attribution windows, and whether signals are person-, device-, or household-scoped.
Run a clean-room or matched-sample test before using the graph for campaign lift.
Document refresh cadence, join confidence, and permitted measurement outputs in the order form.
A pre-computed similarity-vector dataset that lets buyers measure how close two records are in identity-attribute space. Used for matching, deduplication, cluster analysis, lookalike modeling, and identity-graph QA without having to compute the distances themselves.
Who uses this product?▾
Data-engineering teams maintaining identity graphs, modeling teams building lookalikes, and fraud / risk teams running link analysis. It's typically licensed alongside the MAID Feed and the consumer file as an acceleration layer rather than as a standalone product.
What's the refresh cadence?▾
Weekly. The vectors are recomputed against the latest consumer file and identity graph each week; deltas are shipped so buyers can update their downstream stores without reprocessing the full dataset.
How is it delivered?▾
Parquet on S3 by default, with custom dimensions (which attributes are folded into the distance computation) configurable per buyer. Ask sales for the methodology brief before licensing, the configuration choice meaningfully changes downstream model behavior.
Is sample data available?▾
Yes, a sample slice with synthetic identifiers is shared under NDA so engineering teams can validate format, distance distributions, and integration paths before signing the MSA.
✓ Opt-Out Request Honored via Global Privacy Control
We use cookies and similar technologies to improve your experience. No marketing cookies are pre-selected. Learn more in our Privacy Policy. Your Privacy Choices.
