> For the complete documentation index, see [llms.txt](https://docs.dstillery.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.dstillery.com/data-and-audience-products/data-and-audiences/dstillery-data-and-ds-1-how-they-fit-together.md).

# Our data & DS-1: how they fit together

Explore the consented, scaled, and specialized signals behind Dstillery audiences, and how DS-1 turns them into activation.

{% hint style="success" %}
Dstillery combines consented user journeys, broad website activity, and specialized partner signals. One layer teaches intent. The other finds it at scale.
{% endhint %}

### The two core data layers

{% columns %}
{% column %}

#### 1. Opted-in panel data teaches intent

About 2 million fully consented users reveal the journeys behind consumer decisions.

Their cross-site visits, actions, and behaviors train the model.

This data is used strictly for model training. It never targets people directly.
{% endcolumn %}

{% column %}

#### 2. Website visitation data reaches at scale

About 400 million devices capture web and mobile behavior across billions of daily events.

This data comes from bidstream partners, SSPs, exchanges, and web publishing tools. These include ad, commenting, and sharing widgets.

The model uses this dataset to find devices showing learned intent patterns.
{% endcolumn %}
{% endcolumns %}

**Panel data teaches intent. Website activity makes it actionable at scale.**

For a travel campaign, panel journeys might reveal future bookers comparing reviews before checking flight options. Website activity then identifies devices showing those pre-conversion behaviors.

### Specialized enrichment signals

Specialized partner inputs add category-specific depth to the core behavioral engine.

| Specialized signal | Primary use case        |
| ------------------ | ----------------------- |
| NielsenIQ/GfK      | CPG and auto            |
| PurpleLab          | Healthcare              |
| ScreenEngine       | Media and entertainment |
| Resonate           | Consumer insights       |
| Emporia            | B2B                     |

The underlying model and daily refresh cycle remain the same across specialized inputs.

### The multimodal AI connection

These data layers form the behavioral foundation of Dstillery's multimodal AI. The model combines behavioral journeys with language-derived insights and specialized partner signals to understand intent more completely.

Learn how these signals work together in [How our multimodal AI works](/data-and-audience-products/data-and-audiences/how-our-multimodal-ai-works.md).

### From signal to activation

Start with the seed signal you already have:

* First-party data or CRM lists
* Brand or competitor domains
* Search terms or keywords
* In-market behavioral signals
* Retail purchase intent

Then the model activates it:

{% stepper %}
{% step %}

#### Learn the journey

The model observes sequential behaviors. It distinguishes genuine consideration from casual browsing.
{% endstep %}

{% step %}

#### Recognize similar intent

Website visitation data identifies devices across the open web showing comparable patterns.
{% endstep %}

{% step %}

#### Deliver the audience

The audience is packaged for DSP activation. Each device is rescored every 24 hours.
{% endstep %}
{% endstepper %}

### Where this shows up

This workflow powers Dstillery's core audience products:

* [Custom AI audiences](/data-and-audience-products/audiences/custom-ai-audiences.md)
* [Custom built audiences](/data-and-audience-products/audiences/custom-built-audiences.md)
* [Custom search lookalikes](/data-and-audience-products/audiences/custom-search-lookalikes.md)
* [Retail purchase intent audiences](/data-and-audience-products/audiences/retail-purchase-intent-audiences.md)
* [Pre-built audiences](/data-and-audience-products/audiences/pre-built-audiences.md)
* [CTV-optimized audiences](/data-and-audience-products/audiences/ctv-optimized-audiences.md)
* [Custom patient targeting](/data-and-audience-products/audiences/custom-patient-targeting.md)

Some products use specialized partner signals more heavily, depending on the category. The underlying model and daily refresh cycle remain the same.

### Where DS-1 comes in

DS-1 turns this intelligence into an audience on demand. Start with an objective, a seed, or an existing audience.

Build it yourself in DS-1, or bring your goal to the Dstillery team. Both paths use the same platform, underlying model, and scoring workflow. DS-1 gives you direct control over the workflow.

### Data principles

{% hint style="success" %}
**Built for your brand.** Each campaign uses a custom model, not a generic shared audience pool.

**Current by design.** Custom audiences refresh and rescore every 24 hours.

**Tuned for scale.** Audiences are ranked against Dstillery’s full device universe to optimize campaign outcomes.
{% endhint %}

### Responsible data use

Dstillery models behavioral intent to target patterns, not known individuals. See the [Dstillery Privacy Policy](https://dstillery.com/privacy-policy/) for data handling and compliance practices.