Create an Optimizing (Auto-Targeted) Campaign

The Optimizing Campaign delivers the most effective experiences at any point in time by automatically distributing traffic to the best performing experiences.

Based on a proprietary optimizing model, Optimizing Campaigns allow marketers to use an automated approach to improving site performance. With an Optimizing Campaign, fewer visitors see under-performing experiences. As visitors respond to different experiences over time, more traffic is automatically sent to the best performing experiences. Visitors who fall in marketer-defined segments are sent to the best performing experience for that segment.

When interpreting the Summary Report, ask the question: "Is my self-optimizing campaign performing better or worse than if I showed random traffic (each experience evenly) to all visitors?"

Look at the Insights Report to understand relationships between certain segments and experiences, according to the rules generated by the algorithm.

Note: An optimizing campaign cannot be optimized for engagement metrics.

This campaign type is appropriate if:

  • You want automation, while still maintaining control over the experience creative
  • Your Web site tends to attract segments with significantly different behavior

Unlike A/B campaigns where a visitor has the same experience until he converts, optimizing campaigns might serve a new experience per visit (or session). You should be aware of this fact if your campaign goal requires users to see a consistent experience across visits. Conversion is attributed to the experience served at that time. When viewing reports, it is most appropriate to select "Visit" as the counting methodology.

If you've defined segments for your campaign and a visitor matches multiple segments, the algorithm serves the visitor the best performing experience for just one of those matching segments, rather than a combination of segments. You can define compound segments like:

  • Windows and Firefox
  • Windows and IE
  • Mac and IE
  • Mac and Firefox

However, it's best to use the Optimizing Campaign with a few distinct segments rather than many granular segments. A rule of thumb is five segments. However, sites or pages with more traffic might accommodate more segments. For optimum performance, consider using a 1:1 campaign, which uses even more sophisticated modeling approaches to automatically target each individual by accounting for everything known about that visitor.

To creating a self-optimizing campaign:

  1. From the Campaigns menu, select Create > Optimizing or click Self-Optimizing on the Campaigns List page.
    Note: This feature must be enabled by your account representative.
  2. Select the target metric:
    • Conversion Rate
    • Revenue Per Visitor
    • Average Order Value
    Note: Optimizing campaigns cannot be optimized on the engagement KPI.
  3. Create the experience and load offers following the steps for an A/B or Multivariate test or campaign including setting up targeting, segment filtering, and/or success metrics analysis as needed.

    In addition to directing traffic to the best performing experience at any given time, the self-optimizing campaign algorithm takes marketer-defined campaign segments into account. For example, if you define two segments, "IE browser" and "Firefox browser," the algorithm examines performance data for these segments and generates traffic serving rules like "Visitors with IE browsers should see Experience A, while visitors with Firefox browsers should see Experience B."

    We recommend you define segments that are meaningful for your business. Other potentially useful segments include "visitors from Affiliate A," "visitors from Affiliate B," "visitors from paid search," and "visitors from organic search."

  4. Alternatively, you can convert an existing A/B campaign into a self-optimizing campaign by clicking Automate at the top of the campaign pages.

    Important: The campaign statistics associated with the original campaign are lost when you click Automate. To keep a history of your existing campaign with its reports, consider first copying the campaign and then auto-targeting the newly copied campaign.