Creating an Audience

The first step to creating an audience is to click the ‘Create New Audience’ button on the top right of Audience dashboard.

When you click the button a wizard will be displayed as a dialogue over the main Audience screen where you can input the:

• Audience Name: Choose a name that will allow you and others in your organisation to easily identify the audience from a list and in DSP Seats or other external tools.
• Date range: This is the time period during which a profile matching the query will have been active on the network. There are a number of set date ranges - the last 7, 14, or 30 days, or the current month - or a rolling audience that runs every 24 hours with a 30 day lookback. To initiate a rolling audience query turn the ‘Rolling’ switch on, noting that a rolling audience will only expire if deleted or if the user sets an expiration date. To set an expiration data turn the ‘Expiration Date’ switch on and use the date dropdown that appears.
• Audience Attributes: This is your chance to add the desired attributes and logic rules you require to define your audience such as sites, numbers of visits, interests, probabilities, etc. See more on attribution selection below.
• Audience Extension: This provides the ability to extend audiences within households by identifying other devices within the same household.

Attribute selection

There are a number of attributes that can be used to create audiences, either in isolation or in combination. The following all assume the logic rule “is” but you can also outline attributes to omit by using the “is not” logic rule:

• Property: The websites used as a source for your query. Adding a property means only profiles who visited that property during the selected time period will be returned. You can also assign a ‘minimum visits’ rule to seek more engaged audiences.
• URL path: This returns only users that visited a specified URL within the organisation’s domains.
• Referrer: This string match returns profiles that arrived into the network from referrers that CONTAINS or DOES NOT CONTAIN the value added to this field.
• Device: The source device used or not used by a visitor. This could be Mobile, Desktop, or Tablet.
• Location: This returns only users from or not from the matching geographical area.
• Custom events: Where custom events have been tracked (e.g. users that did or didn’t subscribe) you may want to target these, this attribute enables you to target users relating to those custom events.
• Intent: This returns profiles which have or don’t have the matching intent or interest based signals that were active during the selected time period. See below for further details on our data scoring.
• Brands: This allows you to find audiences with/without an interest in specific brands. Similar to intent scoring, you can see the level of brand intent - see below for more details on data scoring.
• Demographics: This allows you to include or exclude a vast range of different demographics including gender, household configuration & ownership, relationship status, career, technographics, social grading, ethnicity, age and income levels. In addition, you can set confidence levels.
• Prebid: This allows you to build audiences from specific prebid campaigns or those that have generated CPMs within set parameters.
• Keyword: This allows you to build audiences from specific keywords associated with content being consumed.

You should also note there's the option to use 1st party data only or both 1st and 3rd party data, simply by setting the relevant toggles below the attribute selection dropdown.

Data scoring

Scoring our data provides confidence to our users when they're building audiences, which is why when building audiences using Intent or Brand Affinity they'll see a slider module such as below:

Users can filter audiences based on the level of intent for taxonomy nodes in order to target broad or niche audiences. Carbon's algorithms calculate intent scores based on 4 core statistical features of a user's site visits: frequency of intent, recency of intent, periodicity or regularity within a given intent category, and source & site type such as whether the site was an ecommerce site or general interest site.

Carbon also enables the filtering of audiences based on their brand affinity in order to target those with varying levels of awareness and advocacy for specific brands.

Logic operator

Often you will want an audience to be more sophisticated than is possible with a single filter - this is where the AND/OR logic rules come in.

If you are aiming to combine multiple filters (i.e. “this property with this interest”) then you’ll need the AND operator. Click the ‘Add rule’ button and you can start adding multiple filters to the group to make it as specific as you require. You can, of course, combine the AND/OR logic rules to create even more powerful audience groups. Just be mindful that the more specific your audience is the smaller the pool of potential profiles will be.

When you have finished building your filters and logic, click the ‘Create Audience’ button and you will be taken back to the Audience overview page. Querying can take several minutes depending on the server load and query complexity. However there is no need to keep the page open as the query will complete in a background process and show in the UI when available. A notification will also be displayed when the query has completed.

Real-time reach estimates

Realtime reach estimates enable faster and more accurate audience building that give users confidence in the audiences, as well as increasing the speed to activation. As a user adds attributes to their audience, the 'Estimated audience reach' figure will change to give an accurate representation of the audience size.