Understanding Different Delivery Schemes

For advanced use cases, Seventh Sense has developed a number of delivery schemes for testing purposes.  This knowledge base article provides insight into how and when to use them.

During the scheduling process, to select a delivery scheme other than the default, you can use the drop down menu next to "Delivery Scheme" similar to the below.

Delivery_Schemes.png

Below is a listing of the delivery scheme types that are options when scheduling a mailing.

  • 100% Personalized
  • 50% Personalized / 50% Account Top Time
  • 100% Randomized
  • 100% Randomized (Even Distribution)
  • 100% Randomized (Weighted by Account Average)
  • 100% Personalized, Drop Missing Profiles
  • 100% Randomized, Drop Found Profiles

100% Personalized

This is the default setting and what is used as the most common method of scheduling a mailing.  When this scheme is used, the system will personalize the email scheduling time for any person on the list that has a profile and use a weighted randomization time for any person on the list that does not have a profile.

The below example shows the distribution of a list for both how many people were predicted and randomized during a sample 24 hour delivery window.

100_Percent_Personalized_-_Full_List.png

If we toggle off "Random Scheduled" we can see the distribution of how many people with predicted times will get their email during each hour in the delivery window. 

100_Percent_Personalized_-_Predicted.png

If we toggle off "Predicted Scheduled" we can see the distribution of how many people with randomized times will get their email during each hour in the delivery window.  Additionally, you can see how the randomized component will weight towards heavier hours of engagement in your overall email program.

100_Percent_Personalized_-_Random.png

50% Personalized / 50% Account Top Time

In this delivery scheme, the system will use two algorithms when creating the scheduled deliveries.  First, the system will automatically split the list / audience into two equal parts (not only by number of people, but also amount of engagement and last period of engagement). 

For half of the list, it will use the 100% Personalized method which is reviewed in the section above.  For the second half of the list, the system will pick the peak engagement time for your overall audience within the scheduled delivery window (also known as an Account Top Time) and use that as a mechanism to more or less blast the email at this time. 

While this method of scheduling was designed to test the effectiveness of personalized vs. blast times, the backend analysis requires professional services at this time and is not the recommended way of accomplishing this. 

50_Percent_Personalized_-_50_Percent_Account_Top_Time.png

If we toggle off "Top Time Override Scheduled" and "Top Time Scheduled" we can see the distribution of how many people were scheduled using the 100% Personalized method.  

50_Percent_Personalized.png

If we toggle off "Predicted Scheduled" and "Random Scheduled" we can see the distribution of how many people were scheduled using the Account Top Time method.  

50_Percent_Account_Top_Time.png

100% Randomized

In this delivery scheme, the system will pick a completely random time for each person in the list to be delivered their email within the scheduled delivery window.  This can be useful with ABM strategies where you're sending to a large number of people within a single corporate domain.  It can also be useful to remove potentially older bias data caused by months or years of blasting email at the same scheduled time.   

100_Percent_Randomized.png

100% Randomized (Even Distribution)

In this delivery scheme, the system will pick a random time for each person in the list to be delivered their email within the scheduled delivery window similar to the above, however, will distribute the randomization more evenly over the scheduled delivery window.  This can be useful when testing to find a top engagement time for a particular list, similar to the study we wrote about in this article.  It can also be useful with ABM strategies where you're sending to a large number of people within a single corporate domain or to remove potentially older bias data caused by months or years of blasting email at the same scheduled time. 

100_Percent_Randomized_-_Even_Distribution.png

100% Randomized (Weighted by Account Average)

In this delivery scheme, the system will pick a random time for each person in the list to be delivered their email within the scheduled delivery window similar to the above, however, will weight the number of people scheduled based on the overall engagement of your entire audience.  This is the systems default mechanism of using randomization throughout the system as we believe it provides an even balance of randomization and taking advantage of data you have on your overall audience.

100_Percent_Randomization_-_Weighted_by_Account_Average.png

100% Personalized, Drop Missing Profiles

In this delivery scheme, the system will only schedule people in a list that have a profile in the system.  This can be useful for testing how people with profiles engage vs. people that have no profiles.  This test is typically used in addition to 100% Randomized, Drop Found Profiles method described below.

100_Percent_Personalized_-_Drop_Missing_Profiles.png

100% Randomized, Drop Found Profiles

In this delivery scheme, the system will only schedule people in a list that do not have a profile in the system or that do not have any engagement data that can be used to predict a time.  This scheme uses the 100% Randomized method of scheduling people.  This can be useful in testing how people with no profiles are engaging with your email program and also be used in conjunction with activity classes to suppress a smaller portion of a specific activity class, such as your Passive and Inactive audience that have never engaged. 

100_Percent_Randomized_-_Drop_Found_Profiles.png

Have more questions? Submit a request

0 Comments

Article is closed for comments.
Powered by Zendesk