Use time constraints in segment definitions

Transcript

In this video, we’ll cover how to create audiences for time sensitive use cases in experience platform. Specifically, we’re going to use a segment builder in the platform interface to build some time constrained audiences to meet the requirements of some example, use cases before some background. Segment definitions are one of the ways you can build audiences in platform. A segment definition is a set of rules that can take profiles from your main population and group them into smaller addressable audiences based on shared traits and behaviors. In practice, these rules are often bound to specific time dimensions, such as customers who have a birthday this month, customers who viewed a product in the last seven days but have not purchased anything in over a year and so on. This is where time constraints come in. In addition to segmenting audiences based on attributes or events, they have in common, you can also segment based on when they happened and in what order. You can apply all kinds of different time constraints at different levels of your segment logic. But rather than list them out here, let’s jump into the segment builder and build out some examples. We’ll start with a simple use case. Say we want to build an audience of all customers whose birthday is today. So we can send them a personalized birthday greeting. We can achieve this by using a time constraint applied to the customer’s birthday, which happens to be a time based attribute in our profile data. So in the left rail under attributes, we’ll search for birth date and we’ll drag and drop that into the canvas. You can see by default this attribute rule is set to today with the ignore year option Disabled today works for our use case, but since we want to segment based on the person’s birthday rather than their birth date, we’ll make sure to enable the ignore year option. So we’re only checking the month and day. That’s all we need to do for our first segment. So now let’s try something a bit more complicated for our next example for this audience. We want to capture power users that have made at least five purchases this year. Since a purchase is an event, we’ll click into events and search for the checkout event type. When we add the event to our definition, a time constraint appears on the canvas, which is set to any time by default. This is called a canvas level time constraint. In order for a profile to qualify for this audience, they must pass all the event criteria we put within the canvas, within whatever time constraint we put here. If we were using multiple events like we’ll show later, this constraint would apply to all of them collectively. Since we’re only using one event type, though, we can set the canvas level constraint to this year. From the list of options, and that’ll ensure we’re not using purchases from previous years in our count. Now we just have to adjust the number of purchase events we want in that time period. So we’ll click into the event card for checkout and under event rules will use the frequency dropdown to set this card to at least five. That’s our second segment done. Just one more to go for this last audience. We’re going to be targeting a recent card abandoned ours. Specifically, we want this audience to capture all customers who added a product to their cart in the last 24 hours and haven’t followed up with a purchase for more than 6 hours after their last add to Cart event. Breaking this down, there are two event types that are relevant for us an add to CART event and a purchase event or the lack of one in this case. Let’s start with the add to cart event. We’ll search for product list adds and when we add it to our definition, the canvas level time constraint appears again. We’ll keep it set to the default. Next, we’ll need our purchase event. So just like before, we’ll search under events, this time pulling up checkouts. Since we want to make sure we’re checking that purchases didn’t happen after the CART add event. We need to be careful where we drop the second event in the canvas. If we drag the new event over the card for checkout, one of its edges highlights to tell us where the new card will be placed once we drop it. Placing it on the left would cause this rule to check if this new event happened before the cart. Add the right side would check if it happened after and the top or bottom would disregard order entirely. There are tons of use cases for all of these options, but for hours we’re going to drop this event to the right. Once added, we have this arrow here that indicates checkout needs to happen after product list adds. You can also see our canvas level constraint has expanded over both event cards since it applies to the whole sequence. Before we continue, we’ll click into the checkout card and switch the setting to exclude. So now we’re checking that this event did not happen as part of our criteria for this audience, since we want to target customers who added to their cart in the last 24 hours, we can add a time constraint to that event card specifically rather than the whole sequence. Click into the product list adds card, then open the options menu under event rules. When we click Apply Time Rule, we get some additional controls to set a constraint for this card. Now what counts is the last 24 hours not really changes as time moves forward. So we’ll use the rolling range option for this constraint and set the range to 24 hours ago. Next, we need to add a second time constraint, which represents that six hour period before a stale card is considered abandoned. To do that, we can click the clock icon above the sequence arrow between the cards, choose after and will enter six for the value and switch the units 2 hours and that’s it. We have a definition that uses three levels of time constraints. The sequence of events itself. Our card level constraint for product list adds a a constraint for the time between cards. If you followed along, you may have noticed that the segment builder gives us a plain text description of our definition that automatically updates as we make changes in the canvas. It’s good practice to check this description before you create your segment to confirm you’ve set up your logic correctly, especially when the segment uses multiple time constraints like this one. So according to this, we have an audience that has at least one product ads event in the last 24 hours and then after 6 hours did not have any checkout events recorded, which is exactly what we’re looking for. As a quick recap, here are all the types of time constraints we’ve covered in this video and where they’re configured in the canvas. At the attribute level, there are time based attributes. These are customer attributes that use a data time as their data type like birth date. You can figure these like any other attribute in the canvas at the event level. We went over multiple constraint types. The canvas level constraint applies to the entire sequence of events in the segment definition. This appears as a dropdown list above all event cards in the canvas sequence constraints to find whether multiple events need to happen in a specific order and what the timing between those events needs to be. These are created by placing event cards horizontally next to each other, then using the clock icon above the sequence arrow. Finally, card level constraints apply to a single event in the canvas under the event rules section that appears when you click the card. Enable the apply time rule toggle in the sections options menu to access the appropriate controls. So that was a high level intro to the kinds of time constraints you can add to your segment definitions and platform and the kinds of audiences that can be used to capture for advanced marketing use cases. Thanks for watching.