Creating a New Experience with Dataset Targeting
You can use a dataset to create targets for the WHO portion of an experience. Each dataset field has a set of options and operators that are defined by the data type that you specified during the upload processes. You can use any and all of the uploaded datasets to personalize an experience for site visitors. Follow these steps to accomplish this goal.
- In the top navigation bar click EXPERIENCES and then Web.
- Give the experience a name, and then complete the WHY settings.
- Click WHO, click ADD TARGET, and then click Datasets.
- Click the attribute (column) from the specific dataset that you want to use.
- Select an operator and then enter a value, if required. See Dataset Target Types in this documentation for more information.
- Repeat steps 4 and 5 as necessary to add the dataset targets you want to the experience. See Combining Dataset Targets in this documentation for more information.
- Complete the WHAT and WHEN settings, and then click PREVIEW or ACTIVATE.
Dataset Target Types
You can use dataset targeting for any column, or attribute, within a dataset. Each attribute is associated with one of four target types:
- Text targets are made up of text values, such as
female. Their operators include Equal to, Contains, Starts with, and Ends with.
- Number targets are made up of numbers, such as
7000. Their operators include Equal to, Greater than or equal to, and Less than or equal to.
- Date & Time targets are made up of dates and times, such as
10/11/20 at 12:00 AM. Their operators include Before, After, Within next days, and Within last days.
- Boolean targets are true/false statements. They don't have operators but instead have a toggle that you can switch from TRUE to FALSE.
See Customer Dataset Upload Requirements for more information about allowed data types and other aspects of a dataset to ensure the files you upload are formatted correctly for use as WHO targets.
Combining Dataset Targets
You can combine different customer dataset targets to reach your intended target audience for every experience.
Target a Group in a Dataset
You can configure a customer dataset target to show an experience to everyone who matches an ID contained within a dataset and who is part of a specified group as captured by an attribute in that dataset.
For example, you create an experience that targets actively engaged members of your site's rewards program. You then configure the WHO settings using the customer dataset that contains the IDs of program members and the engagement status of each one, among other datapoints.
Target Everyone Except One Group in a Dataset
Just as you can configure WHO settings to target a specific group within a customer dataset, you can also exclude a specific group from the target.
For example, you create an experience that targets rewards program members who've spent at least $200 over the lifetime of their membership. Because you want to exclude anyone who has spent less than $200, you must switch the inclusion toggle on the configuration panel from INCLUDE to EXCLUDE.
Use Multiple Dataset Targets
All defined parameters within a single target are treated as "or" statements. For example, if you want to create an experience targeting customers within a customer dataset who are located in Oregon or California or Pennsylvania, then you only need to define a single dataset target that includes all three states as values.
Sometimes you must create multiple targets and then associate them using the AND/OR toggle.
Building on the previous example, you can further specify the targeted audience as customers within the customer dataset who are located in Oregon or California or Pennsylvania and who have a buying propensity of at least 50 or higher. You can achieve this goal by first creating a second target based on the
propensity score attribute from the same dataset. Configure that second target using the Greater than or equal to operator and
50 for the value. Once you save this second target, ensure that the AND/OR toggle is set to AND.
If you want the experience either to target customers found in a customer dataset who are in Oregon or California or Pennsylvania or to target customers in that same dataset who have a buying propensity score of 50 or higher, simply change the toggle to OR.
Include vs Exclude Toggle for Values
The INCLUDE/EXCLUDE toggle that appears on the target configuration panel can have a big impact on an experience when you use customer dataset targets. When the toggle is set on INCLUDE, the experience appears for all customers who are in the dataset with the selected attribute and who match the specified target value.
Conversely, when the toggle is set to EXCLUDE, those customers who meet the target criteria are excluded from seeing the experience, which does appear to customers who are in the dataset but who do not match the specified target value.
You can also target all customers who are in a customer dataset but have no value whatsoever for the attribute you're targeting. For example, you want an experience to target customers in the rewards program dataset for whom you haven't captured a program membership status, which appears as
N in the attribute column if captured. To achieve this goal, configure the target with the Has value operator, and then set the toggle to EXCLUDE.
If you accidentally target all members of a customer dataset or no members of a customer dataset, one of two warning messages appears.
Consider this example. When creating an experience, you want the WHO targets to use a customer dataset that contains the
channel of last order attribute, the value for which is either
catalog. You build one target to exclude customers within the dataset whose last order came from POS. You build a second target to exclude customers within the dataset whose last order came from the catalog. You then set the AND/OR toggle to OR. Upon switching the toggle, a warning message appears:
Warning! With this combination of OR targets, all visitors who are in your dataset will see this experience because each rule is evaluated independently. Are you sure that's what you intended?
Incidentally, if you had built a single target to exclude customers within the customer dataset and had input
catalog as the values, no warning message would appear as a result.
Consider another scenario that would result in a warning message. The same attribute of the same customer dataset used in the previous example is the basis again for two WHO targets. This time, customers within the dataset whose last order came from POS are the target for one of them. Customers within the dataset whose last order came from the catalog are the target for the second one. You then set the AND/OR toggle to AND. Upon switching the toggle, a different warning message appears:
Warning! With this combination of AND targets, no visitor to your site will see this experience because mutually exclusive rules cannot both be true. Are you sure that's what you intended?
If you'd built a single target to include customers within the dataset and had input
POS as well as
catalog as the values, no warning message would appear because the two values within the single target are treated as "or" statements.
When determining whether to configure target criteria to specifically exclude from the experience those customers who match that criteria, keep in mind the following implicit exclusion logic:
- If a visitor has not been identified and thus doesn't have an ID, they're excluded from the experience.
- If a visitor has been identified, but the targeted dataset has no matching row for that person, they're excluded from the experience.
- If a visitor has been identified, the targeted dataset has a matching row for that person, but no value appears for the targeted attribute, they're excluded from the experience.
Sample Size Expectations
You should consider several factors to get the most out of customer data onboarding and thus WHO targets that use customer datasets.
The more pervasive an on-site identifier is on your site, the larger the initial sample size is. Conversely, an on-site identifier that's less pervasive—for example, on-site identifiers located only on a few pages or on pages only accessed a few times over the course of a day—produce a smaller initial sample size.
Furthermore, under most circumstances only a percentage of site traffic identified by the on-site identifier qualifies for an experience. This diagram shows the potential audience funnel for an experience using customer dataset targeting.
Before using a customer dataset for a WHO target in an experience, consider running several Do Nothing experiences to compile data and increase the potential audience pool eligible for the on-site identifier. Doing so gives you an idea of what percentage of site traffic contains an on-site identifier.
Use your knowledge of your customer base as well as the data you collected in previous experiences on your site to determine whether these site visitors represent a high-value segment. If the sample size for your audience is too small or doesn't represent a segment of high-value customers, consider revising the experience and shifting the focus to a wider or more general set of targets.