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
Start here (overview)
If you’re evaluating identity resolution for onboarding or measurement, begin with the canonical GSDSI identity graph page and then validate the procurement model in the linked resources.
Procurement checklist
- Match rates: ask for cohort/geography-specific match rates, not global averages.
- Refresh: confirm cadence and decay dynamics for MAID/HEM freshness.
- Governance: permitted uses, retention, opt-out/delete SLAs.
- Delivery: Snowflake/BigQuery/S3/API + Parquet/CSV/JSON.
Related products
Related solutions
How to evaluate an identity graph (without getting misled)
Most identity contracts fail for one reason: buyers validate the wrong thing. The right pilot is a small matched sample that proves join yield in your geography and cohorts, plus governance that fits your activation and measurement stack.
Ask for these artifacts
- Match-rate readout: by cohort/geo, not a single global number.
- Refresh + decay: what changes daily vs weekly vs monthly.
- Join keys: exactly which identifiers you receive and how they’re normalized.
- Governance: opt-out/delete SLAs and retention boundaries.
Red flags
- “100% coverage” claims without cohort breakdowns.
- No written handling rules for downstream deletion propagation.
- Evaluation that skips dedupe rules and householding assumptions.
- Ambiguous “compliance” language without permitted-use boundaries.
Fast path (2 weeks)
- Start with a matched sample + target join yield.
- Confirm delivery path (S3/SFTP/Snowflake) + formats.
- Lock governance terms before scaling volume.
- Then move to a production license.
Related buying guides
If you’re evaluating adjacent datasets, start with the relevant hub first to avoid overlap and ensure you’re citing the canonical routes.
Vendor comparisons
If your stakeholders are asking “GSDSI vs ___”, use the comparison pages as a pilot rubric and procurement memo template.
Frequently asked questions about identity resolution procurement
When should I use this buying guide versus the identity graph product page?
Use /identity-graph for product specs, schema fields, and MAID feed coverage. Use this /identity-resolution hub when procurement needs a checklist: match-rate evidence, refresh and decay, governance questions, and links to diligence resources before signing.
What match-rate evidence should I request from any identity vendor?
Ask for join yield on your own hashed seed, broken out by geography, cohort, and confidence tier, not a single deck average. Request last_seen distributions, opt-out propagation proof, and a change-delta sample so stale edges can be filtered in warehouse merges.
How often do MAID→HEM linkages refresh?
US MAID→HEM programs on GSDSI refresh on a daily cadence; international pairs follow product-specific schedules. Document decay expectations using last_seen fields and the device graph decay resource before scaling activation volume.
Which identifiers should be in scope for a pilot?
Typical scopes include MAID, SHA-256 HEM, household or device cluster IDs, CTV/ACR IDs where TV joins matter, link_type (deterministic vs probabilistic), and confidence scores. Align the field list to your DSP, CDP, or measurement stack before the matched sample ships.
Should we run a pilot before a production identity license?
Yes. GSDSI recommends a matched-sample pilot (10K-100K seed rows) with documented match-rate targets, deletion handling, and delivery path validation. See /pilot-process for the standard enterprise checklist.
Where can I compare GSDSI to other identity vendors?
Use /compare/gsdsi-vs-liveramp and the /comparisons index as diligence memos. They spell out scope differences and what to validate in a side-by-side pilot rather than marketing claims.