| Audience | How you’ll use this article |
|---|---|
| Marketing teams | Understand how identities are unified and how resolved profiles power audiences, personalization, and activation. |
| Data teams | Learn how identity graphs are built, validated, and exposed through warehouse tables and Customer Studio. |
| Platform admins | Follow the recommended setup and validation flow to safely introduce Identity Resolution. |
Overview
This guide walks you through the end-to-end process of implementing Identity Resolution from understanding how identities are formed, to creating an identity graph, configuring Golden Record, validating match quality, and activating unified identities.
The steps below follow the order most teams use in practice and link out to deeper documentation where needed.
Implementation overview
| Steps | What you’ll do |
|---|---|
| Understand how Identity Resolution works and Deterministic matching | Learn how identity graphs, deterministic matching, identifiers, and incremental runs work together. |
| Prepare your data | Ensure required tables exist as models, validate timestamps, and confirm identifiers are ready for matching. |
| Create an identity graph | Create and configure a new identity graph built from your source data. |
| Configure Golden Record (optional) | Define survivorship rules and create a one-row-per-identity table. |
| Monitor and validate your identity graph | Inspect match rates, resolved identities, and Golden Record values before downstream use. |
| Use identity outputs | Query identity tables, model unified data, and understand when to use each output. |
| Set permissions | Control who can view or edit identity graphs by managing source permissions. |
Step-by-step guidance
1. Understand how Identity Resolution works
Start by learning the mental model behind Identity Resolution so you know what to expect from setup and results.
This step explains:
- What an identity graph is
- How deterministic matching works by default
- How identifiers, priorities, and limits affect grouping
- How incremental runs impact identity stability
→ See How Identity Resolution works
→ Learn about Deterministic matching
2. Prepare your data
Before creating an identity graph, your data must be ready.
Identity Resolution only works with models defined in Hightouch—it does not operate directly on raw warehouse tables.
In this step, you’ll learn:
- Which tables need to exist as models
- Timestamp requirements for incremental processing
- How identifiers should be structured and normalized
- Common data readiness checks
Each table used in Identity Resolution must already exist as a model. Create or update models before starting the identity graph setup.
→ See Prepare your data
3. Create an identity graph
Once your data is prepared, create an identity graph using the Add identity graph workflow in the Hightouch UI.
This step walks through:
- Selecting a source
- Selecting and configuring models
- Assigning identifiers
- Defining identifier priority and limits
- Choosing an output prefix
- Running the graph
This establishes how identities are formed and grouped.
→ See Create identity graph
→ Learn about Identifier rules
4. Configure Golden Record (optional)
After your graph is defined, configure Golden Record to create a flattened, canonical profile table.
Golden Record allows you to:
- Select which identifiers and traits appear in the unified profile
- Define survivorship rules (for example, most recent, most frequent, source priority)
- Generate a one-row-per-identity table in your warehouse
- Automatically create a parent model in Customer Studio (if enabled)
Golden Record does not affect matching—it controls which values are surfaced after identities are formed.
Survivorship rules can materially affect downstream activation.
For high-impact fields such as email, phone, or CRM IDs, align on selection logic before syncing to external tools.
→ See Golden Record
5. Monitor and validate your identity graph
After running the graph, review results before using identity data downstream.
This step helps you:
- Evaluate match rates and identifier coverage
- Detect superclusters or unexpected merges
- Inspect resolved identities
- Validate Golden Record values
Validation is strongly recommended before enabling scheduled runs or activation.
→ See Monitor and validate identity graph
6. Use identity outputs
Identity Resolution writes output tables to your warehouse representing the current state of your graph.
In this step, you’ll learn:
- What each output table represents
- When to use
__resolvedvs__golden_records - How to model identity data downstream
- How identity outputs integrate with Customer Studio
→ See Lookup table usage
7. Set permissions
Permissions for Identity Resolution are inherited from source permissions. There is no separate permission system just for identity graphs.
Your ability to:
- View identity graphs
- Create new graphs
- Edit configuration, rules, or Golden Record
- Delete graphs
depends on your permissions for the source (warehouse connection) that each identity graph uses.
→ See Identity Resolution permissions
Activate identities in Customer Studio (optional)
Once identities are validated and modeled, you can activate them across Hightouch.
This may include:
- Building audiences and traits in Customer Studio
- Syncing unified identities to downstream tools
- Mapping appropriate identifiers per destination
→ Continue with:
Advanced and optional steps
After completing the core setup, you may choose to:
- Detect and remediate Superclusters
- Trigger rebuilds using the Profile reprocessing API
Identity Resolution is designed to scale with your data—start with a conservative configuration, validate results carefully, and expand as needed.