Create Criteria in Recommendations
Description
In this video, you learn how to:
- Create criteria
Intended Audience
- Business Practitioner
Transcript
Hi, in this video, we’re going to cover configuring criteria within Adobe Target recommendations. I’m Mike Mineer, a partner training instructor here at Adobe.
By the end of this video, you should be able to create criteria, create criteria sequences and upload custom criteria.
The benefits of this are that you can create custom criteria based on your unique need. You can also use custom criteria in a sequence and use custom parameters in criteria. Now when creating criteria, there are several pieces of information that are needed so the system knows what you want. We’re gonna spend a little bit talking about some basic information, the high level. But then we’ll go into the system itself and go through each of these step by step. So each criteria needs first of all some basic information like criteria name, display title whatnot. They need a data source.
Then you need to decide which content do you want to display and this gets into what if there’s not enough products to fill up the entire recommendation? We’ll talk about that. You can then add some inclusion rules and some attribute weighting. So let’s go ahead and go to Target and create some criteria.
To get to the page to create criteria, you click on the recommendations tab on the top and criteria on the left-hand side. Once there, we wanna click on the blue create criteria box and we’re gonna start by choosing the option, create criteria.
Now we wanna start by filling out the basic information. So we’re gonna give ours a criteria name of popular items to upsell. We’re gonna choose to recommend products.
Once here, we wanna click on the blue create criteria box and choose the top option of create criteria.
Now we wanna start filling out the information. We’ll start with the basic information and the criteria name and the display name. The criteria name is the internal name used to describe the activity whereas the display title is the public-facing title. So we’re simply gonna name ours, popular items to upsell and we’re gonna make our criteria name and display title the same. Now with this criteria, we’re gonna try to upsell the most popular products across the site to people that are already purchasing or looking at products.
The description should be used to identify the criteria in the future and should also include information about the purpose of the criteria so that when you come back to it you know exactly what the purpose of it is. The industry vertical will be based on your site but do keep in mind that which vertical you select will change other options later on in this page. As you can see there’s three different options, we’re gonna simply stick with the retail, eCommerce.
The page type will also determine which criteria is available as you create new criteria on your site because it’s going to look at different key values. In our case, we want ours to be available on the product page.
The home page is there by default and we’re gonna keep it there for now but you can remove it if you want to.
Next we need to choose a recommendation key. This is the item that it’s going to look at to place other recommendations on the site. We’re gonna stick with the current item.
Once we choose our key we need to choose our logic. And there are several items of logic for us to choose from here. I do want to point out that the options below change based on what you pick here. If we choose items with similar attributes notice that our data source went away. We no longer need to choose that.
But if we choose one of these other ones, then we need to also choose a data source. So for the sake of the example we’re using we wanna choose the option of people who viewed this, bought this. That’s our goal and our upsell. Now on our data source, there are several things we need to decide. First of all how far back do we wanna look? We can go anywhere from two days on the low end to two months on the top end. And this is simply how much data is our site going to look at? If we choose the longer amount of time like two months our site’s gonna be slower to update seasonally to seasonal items that may come along. For example come a holiday season of some sort, if we’re set to two months, our site might not start showing those holiday items right away and may cost us some potential sales. So the date range is purely based upon your needs and what you need for your site. The behavioral data source is either gonna be inboxes or Target or analytics. If you choose analytics you simply need to choose a report suite for that data source. But for our example here, we’re gonna stick with inboxes.
The next is content. Content rules determine what happens if the number of recommended items does not fill your design. For example if your design has space for five items but your criteria causes only three items to be recommended you can leave the remaining space empty or you can use backup recommendations to fill that extra space. So option one, enable partial design rendering. That would simply put three recommendations into the five slots as I just mentioned and leave the other two blank.
Or we can choose to show backup recommendations and that could potentially fill in those extra two with other recommendations. And I say potentially more than likely will but as long as we keep the enable partial design rendering there, if there’s not that many products to recommend then it won’t show up but if we do this then it’s going to fill in the backup recommendations to fill in all the slots.
Then we have here our inclusion rules we want to use those for backup recommendations.
And we’ll talk about inclusion rules here in just a minute but that just decides do we want to make sure that they are included on these products just like the other products? And then do we want to recommend previously purchased items or not? Now the inclusion rules allow us to say, okay, if everything that matches the criteria we wanna add another level of checking on that. So one thing we can do is maybe we don’t want to sell items that are too expensive or too cheap, we wanna keep them within a relative ballpark. So say within a range of, 75 to 125% of the price of the product. Whatever it is we can add various inclusion rules to make sure that the items we’re showing are more likely to be bought by our visitors. Also, we can make sure that we have inventory in stock so that if they buy it we don’t have to send 'em an email saying sorry this item’s out of stock. And you can even say that you want a certain number of items to be available before you do it.
And another option you have is you can have other inclusion rules based on any attribute you send in, whether it be an entity attribute or profile attribute, whatever it is, you can include that in your rules. You can even do parameter matching. So if the category is the same, is the value text stored in a parameter? So come down here we look at parameters and we can decide there what we want.
On top of inclusion rules we can also add a filtering rule which is just another type of inclusion rule.
These allow us to add entity attributes, profile attributes, parameter matching or a static filter. So if we do something like a profile attribute it allows us to come down and say, if our category equals a value that we’re passing in one of our profile attributes, then we want that to go ahead and be considered, we want the categories to match. Let’s start this one over again.
Next is we can add filtering rules. These filtering rules allow us to add an additional level of rules of which products we want to show. Any value we pass in through an entity attribute or a profile attribute can be used to match to the key or the product we’re looking at in order to show those products in the recommendation. And then finally we have attribute waiting. This allows us to make it more likely or less likely that a particular product or type of products will show up. So let’s say if the category equals, and in this case we’re trying to upsell a particular product, we’ll say if a category equals shoes, we want it to be up to 25% more likely to show up. And then we’ll come in here and we’ll say, if the category equals shirts, we want it to be 25% less likely to show up. And so it allows us to choose which items we want to more likely show up or be less likely to show up. Notice it doesn’t allow us to include or exclude through attribute weighting. It’s just increasing the likelihood as to whether or not they will show up. So we’ll go ahead and click save there. And we have now created our custom criteria. The next option is to create a criteria sequence.
A criteria sequence has us fill out some basic information just like the last one. Has us choose what we want to do with content. Do we wanna enable partial design rendering or whatnot? Show backup recommendations? But then what it has us do is it has us go choose a sequence order.
So on criteria one if I click add a criteria, I can include the one I just created.
But then what I can do is tell the system after all the items that match that first criteria, I want you to put those through a second level and those then must match the top-selling products from across the site.
Then we can go ahead and save that after we give it a name.
We’ll just call it an upsell sequence.
Go ahead and click save.
And now we have a criteria sequence. And the last option, very similar to the others where you go through, you give it some basic information, choose which content you want it to show. But then you have to download the CSV template and match your data to this template based on the key value. But then you can pass in however many custom values you want up to 1000 to be included in your criteria. So you go ahead and save that. And then upload it based on either an FTP location or a specific URL you just put the URL in there, click save and it will upload those values.
So now you should be able to create a criteria, create a criteria sequence and also upload a custom criteria.
Additional Resources
recommendation-more-help
0f172607-337e-442f-a279-477fd735571f