Using a Media DCR as an advertiser or agency

note

This is the flow if your user has been assigned to the advertiser in the Creating Media DCR step. The agency role has the same capabilities apart from connecting data.

Connect Data - The DATA tab

As an advertiser, the first step is to provision your customer list to the Media DCR. The dataset must contain the following columns:

• Matching ID: This is the identifier that will be matched against the publisher user list. You can check which matching ID format should be used in current Media DCR, click the options menu at the top right of the screen and select View configuration. An individual user may appear in more than one audienceType.
• Audience type (optional): An audience type is a subset of users with common characteristics. In this column, you tag the user that belongs to the same audience type. All analytics dashboards, affinity scoring, and lookalike modeling are done separately for each audienceType. Typically this is the campaign, website, or specific customer list that the user was identified being relevant for. Each audienceType is a “universe” — you would get the same insights and audiences if you ran one clean room per audienceType, but you can run more than one in the same clean room for convenience and comparability. If you only run a single campaign with all of your customer data included, this field may be filled with a dummy value.

The combination of matchingId and audienceType should be unique. Duplicates will throw an error during the data onboarding in this table.

Input Table Example

matchingId	audienceType
e4191e26a5d04a3d8ae9008189a0db7168aba7e8b92944ae90c6a15283e22e2b	Sneaker Campaign
fb3f59dbe51d42f3bf4f826ac5c22f1f13d2d9e464804539ad59417827706bbd	Sneaker Campaign
fb3f59dbe51d42f3bf4f826ac5c22f1f13d2d9e464804539ad59417827706bbd	Athleisure Campaign

It is possible to upload a new dataset or select an existing one. Please follow the data onboarding tutorial in the Datasets section for a detailed walk through.

Gain Insights - The INSIGHTS tab

The insights are available if the module Insights was enabled in this Media DCR creation. Both advertiser and publsiher has access to the INSIGHTS tab.

As soon as data is available from both publisher and advertiser, the insights computation will automatically start. This process may take up to 40 minutes, depending on the amount of data. The result is a dashboard that displays customised analytics, designed to help you understand how publisher segments relate to your first party data.

There are 4 views:

1. Basic: a quick overview of the match rate, demographic overlap, and top segments.
2. Detailed: a deeper view, showing how demographics and segments interact in combination with each other.
3. Comparison: a side-by-side view of different advertiser audiences, so you can see how one compares to another.
4. Explorer: a custom data visualization engine that allows you to look at different slices of data along different axes using dropdowns.

The insights can be used for:

• Planning the campaign: higher effectiveness by understanding the audience segments and their demographics.
• Creating captivating and targeted creatives: by getting valuable insights into the audience's preferences and interests, it enables the design of compelling and highly relevant creative content.
• Picking the right inventory: advertisers can make informed decisions about selecting the most suitable inventory for their campaign.
• Verifying the campaign briefing: ensure that the campaign briefing aligns with the target audience and objectives.
• Planning contextual campaigns: identify relevant contexts and environments where the campaign can be most effective.
• Learning about your advertiser's customer base: get valuable information about the advertiser's own customer base, and a better understanding their behavior and preferences.
• Benchmarking the ex-post results: advertisers can compare the actual campaign results with the insights from this dashboard to evaluate the effectiveness of their strategies and make improvements.
• Benchmarking the attribution results: assess the attribution of conversions and measure the impact of the campaign on various metrics.

By leveraging the insights provided in this dashboard, advertisers can optimize their campaigns, improve targeting, and achieve better results.

A few points to note:

• The insights are calculated based on the overlap between the advertiser customer list and the publisher audience.
• A small overlap ratio does not necessarily mean that the publisher audience is not relevant to the advertiser. It just means that the insights are based on a smaller sample size. By running a Lookalike model, the advertiser can expand the audience to include more relevant individuals.
• If there are not enough users in the overlap (less than 150 users) it is probable that there was an issue with the matching ID. Please check again that the ID is the one agreed between publisher and advertiser, and if hashing was required that it has been applied as expected.
• Affinity ratio is used in several metrics. It is calculated as the ratio of "Share in overlap" to "Share in publisher audience". The higher the affinity ratio, the more likely an individual in the advertiser segment is interested in the advertiser's product compared to a random individual in the addressable publisher audience.
• The segments and demographics are strictly related to the publisher inside the Media DCR. The results are custom for each collaboration, therefore it is not intended to be used in a different context.
• Decentriq consults with publishers to increase data quality and ensure consistency, aiming to provide the best results possible.
• Privacy filtering is applied to the data to ensure that no individual user can be identified. This is done by aggregating the data and only showing insights that are statistically relevant.

If the Show absolute audience sizes setting was disabled for the insights, the charts will only show relative values. Some publishers might choose not to show absolute values to advertisers for privacy reasons.

Create Audiences - The ACTIVATION tab

This can only be seen if either Remarketing or Lookalike were enabled during the Media DCR creation. Both Remarketing and Lookalike are supported in the same Media DCR. You can create multiple audiences and make them available to the publisher. They are all independent from each other.

With remarketing and exclusion targeting modules

When accessing the ACTIVATION tab you will see a preset list of all audiences (the ones present in the Audience type column of your customer list) already available for activation. Each audience corresponds to either the overlap (remarketing) between that audience and the publisher's registered users, or the publisher users that are not in the overlap (exclusion).

Once you have generated the audiences to activate, click on Make available to the publisher for each of them. This way, the publisher can export the ID list from the Media DCR and activate the users using their ad serving systems.

With lookalike module

Select the audience type you want to optimize for, and then use the slider to define the desired balance between audience size. You can see an estimated size, quality score, and ROC score immediately. The smaller the audience size, the more precise the campaign. The bigger, the broader the reach will be.

The quality score and ROC score can help you understand how precise and similar your lookalike audience is to the overlapping seed audience. The quality score is a holistic scores that takes into account the ratio between "true positives", people in the lookalike audience and seed audience, and "false positives", people in the lookalike audience who are not in the seed. It's not important to maximize the quality score, but to find good mix of reach and quality. Advertisers seeking a campaign with a large reach will generally try to have the largest reach possible while still maintaining a high overall quality score.

The ROC (receiver operating characteristic) score measures the "true positive rate" -- what fraction of the seed audience is in the lookalike audience. Most advertisers will not necessary try to max this out at 100%, but rather choose a size where the coverage is high and increasing the reach further is giving diminishing returns. You may also click on the "see graph" button to view the ROC curve directly, including a marker for the current slider position.

When clicking on Generate results, an AI model will take the overlap between your and the publisher's customers expand your audience by finding prospective publisher customers that are the most like your existing ones.

You can also click the Exclude seed audience checkbox when you don't want to include your own customers in the Lookalike audience. This does not affect the quality metrics, which are still evaluated against the seed audience, even though it will not be included in the output if the box is checked.

Once you have generated the audiences to activate, click on Make available to the publisher for each of them. This way, the publisher can export the ID list from the Media DCR and activate the users using their ad serving systems.