Inputs
The signals that matter
Names, postal addresses, emails, phones, date of birth, or a person ID you already received. Email and phone can arrive raw or pre-hashed.
Identity Resolution & Enrichment
Identity resolution answers a simple question: who is this record really about?
Enrichment answers the next one: what would be useful to know about them?
Infinite Audience does both in one workflow. Bring the customer data you already have. We connect it to our identity graph, return a durable Infinite Audience Graph Person ID for strong matches, and add the attributes you choose.
Your recordNames, addresses, emails, phones, DOB, or a person ID
DeterministicExact, verified connections in the graphProbabilisticWeighs every detail when data is messySpatialReal-world geography and commuting distanceResolved personPerson ID, match level, confidence, and enrichment
Each step runs only for records that earlier steps have not yet matched to a specific person.
What you can bring
Section titled “What you can bring”- A single customer record from a form, checkout, or sales workflow.
- A small group of records that needs an immediate answer.
- A full customer file in CSV, JSON, JSONL, or Avro format.
- Customer records already living in HubSpot, Shopify, Klaviyo, or another connected system.
You do not need a perfect file. Clear column names help, and the platform gives you a chance to review how they were understood before matching begins.
Multiple values
Every email, phone, and address
Send all of a customer’s emails, phones, and addresses on one record. Each one is tried as its own match point, so no signal is left behind.
Large files
Ready for your data lake
Upload a large list as up to 50 shards straight from your warehouse or lake export, compressed or not, as one logical job.
How matching works
Section titled “How matching works”Real customer data has nicknames, typos, name changes, recent moves, and work addresses. Every record runs through three matching approaches in order, moving on until it reaches a confident person-level answer or runs out of approaches.
Exact match
Deterministic
Verified links in the graph, with smart handling of nicknames, address formatting variations, and last-name changes confirmed by date of birth.
Messy data
Probabilistic
Weighs every detail together when no exact link exists: typos, unit formats, hyphenated surnames, and a mistyped ZIP code. One hard conflict rules a candidate out.
Spatial engine
Spatial
Uses real-world geography to find the right person nearby when text matching cannot, such as a move across an apartment complex or a work address.
Match levels
Section titled “Match levels”Every match is labeled with what it confirms:
| Level | What it confirms |
|---|---|
| I · Individual | The specific person. |
| H · Household | The home and family the record belongs to. |
| D · Digital | A known email or phone, without a confirmed name. |
| S · Spatial | The person, placed by the spatial engine. |
| A · Address | The physical address only. |
Each match also carries a confidence score. When the engine is not confident enough, it leaves the record unmatched instead of guessing.
The spatial engine
Section titled “The spatial engine”When addresses differ on paper, geography reveals the ground truth. Traditional fuzzy matching drops records when street numbers or ZIP codes change. The spatial engine evaluates physical distance, neighborhood density, name frequency, middle names, and date of birth to locate individuals who moved locally or submitted an office address.
Default
Residential
Finds matches traditional fuzzy matching misses—such as moves between buildings in an apartment complex, local moves, or neighboring ZIP codes.
Workplace
Work
Connects a person you only know at a workplace, such as a lead form or event signup, to their home profile within commuting distance.
Mixed lists
Auto
Checks each address. Commercial addresses get the commuting search, and homes stay in their neighborhood. No need to split your list.
What comes back
Section titled “What comes back”File exports preserve your original columns alongside match diagnostics and the enrichment attributes you requested. Real-time API responses return results keyed by row_id so you can join them back to your original records.
A resolved identity
Strong matches receive a durable, organization-safe Infinite Audience Graph Person ID you can use in future workflows.
Match details & confidence
Matched records include a match_level, specific match_type, and a 0–1 match_confidence score. A direct Person ID refresh preserves the prior match level, which may be empty.
The attributes you chose
Select a ready-made attribute bundle or pick individual attributes from the catalog.
A complete result
File exports keep every source row; real-time API results use row_id to connect back to your source records. Matched records gain intelligence; unmatched records stay visible with consistent derived keys.
Enrichment attribute categories
Section titled “Enrichment attribute categories”- Identity
- Audiences & Segments
- Demographics & Household
- Financial & Property
- Media & Behavioral
- Geography
- Land Context
- Census Demographics & Socioeconomic
- Health
- Environment & Risk
- Economic Indicators
- Brand Proximity
- Neighborhood Lifestyle
- Transportation & Mobility
Match once. Keep the value.
Section titled “Match once. Keep the value.”A resolved identity is more useful when it travels with the customer. Keep the returned Infinite Audience Graph Person ID and Infinite Audience can recognize that identity in later supported workflows without repeating the full matching step.
That makes refreshes faster, makes customer history easier to connect, and lets one good match support work across the platform and your integrations.
The ID is salted for your organization: the same resolved person receives a consistent ID inside your workspace and a different ID in another company’s workspace. Unmatched rows may receive a derived ID for continuity, but derived IDs are not resolved people within IAG.
Person IDs & privacy
Organization-scoped, non-reversible IDs for resolved people, plus consistent labels for unmatched rows.
Record licensing
Resolve a customer once, reuse the identity for 12 months, and keep enrichment current.
Real time or at scale
Section titled “Real time or at scale”Both ways of matching run the same engine with complete matching parity between real-time and file workflows.
Up to 100 records per call
Real time
Built for forms, checkouts, and CRM events. A 100-record call averaged about 10 seconds in testing, which works out to about 100 ms per record.
Thousands to tens of millions
Files at scale
Built for full customer lists. In testing, 1 million records matched in about 8 minutes and 10 million in about 17 minutes.
A typical file workflow
Section titled “A typical file workflow”- Upload the file. Give the match a friendly name so it is easy to find later.
- Review the columns. Confirm the suggested mapping and skip anything that should not be used.
- Choose enrichment attributes. Pick an attribute bundle or select individual attributes.
- Let the match run. Follow progress in the platform activity rail.
- Use the result. Download the enriched file or keep the matched identities in the platform.
Start building
Section titled “Start building”In the platform
Upload, review, and download without writing code.
Real-time API
Match up to 100 records per request from your own systems.
File API
Automate large and sharded file matches end to end.