Leveraging structure to rank content items
Customers expect relevant content. Ranking content items in a list is an effective way to present your most relevant items to customers. Structured content makes that such ranking possible.
Previously, we talked about how the details of a few items can be compared. In this post, we will look at how to rank items according to a specific criterion. While both approaches let users prioritize content, they have a different emphasis.
|Presenting a small set of items to allow several dimensions to be compared||Selecting a small set of items based on one dimension to help readers choose which one to look at in more detail|
With a comparison, the reader may not need to look at more information. With a ranking, they use it to decide what items to look at in more detail. Comparisons present a full content item, while rankings preview a few details about an item but don’t present the whole item.
Help users whittle down items
Ranking helps users when they face what seems like a bottomless list of content items. Because they can’t look at them all, they need a shortlist that they can scan and access easily.
You can structure the content to support the ranking of content items.
There are two basic approaches to ranking:
You can use a system-defined ranking to automate the ranking of items. It’s an effective approach when many people have common goals, but the precise items that match those goals will change over time.
The logic that the system uses to rank items will vary according to the criteria used. In some cases, the system will evaluate an element in the content item to perform the ranking. In other cases, the system relies on data about the item from other sources, such as analytics.
The system will evaluate an input for each content item to rank, and then display an output of ranked items, which will display one or more elements for each content item, such as a title. For example, to rank items according to which are newest, the system would compare these dates and then show the names of content items according to which are more recent.
When most users care about similar priorities, it makes sense for systems to rank the items. Common priorities include locating items that are:
- Newest (input: publication date)
- Highest rated (input: user ratings)
- Most popular (input: analytics data)
- Most similar to the current item (input: taxonomy classification)
Typically, the system displays the name or image of the items in a list. It may also display a detail, such as the number of stars in the case of the highest-rated items. You will want to decide what elements to display in your list of ranked items.
You can offer user-defined rankings to account for differences in individual preferences. The user can specify their criteria to the system, instead of relying on defaults. It’s a first step toward personalizing the content of interest.
Users commonly want to prioritize options according to:
When content is appropriately structured, users can rank items according to the criteria they care about, whether it is cost, convenience, or capability. For example, it is common for people to sort items by price or location.
In some cases, users may have multiple criteria, such as when they are looking for the cheapest option that is available now.
As with system ranking, you will want to decide what elements to show in the list of ranked items.
How it’s done
When content items are based on the same content type, they share a common structure and can be compared easily.
The construction of a content type should reflect the prioritization goals and decision criteria that readers use to select content. If you are unsure what these are, you can prototype different options with sample content and present them to users to elicit their feedback. Readers can indicate what’s valuable to them as they whittle down options and what they’d expect to see so they can choose items to view. Ask readers what they most want to see and least want to see. Those preferences are often the criteria that they use to rank content items.
The structure of the content should take into account input variables and outputs displayed relevant to ranking. Two important questions are:
- Have you included in the content type elements that readers want to rank?
- Have you included elements readers expect to see to support choices about items to view?
Generally, content elements are used to rank items, but sometimes ranking will use metadata or taxonomy terms about content items. The system can fetch items that match the value sought (for text) or fall within the range of the value sought (for numbers, prices, and dates).
Numeric values such as prices are easy to rank in ascending or descending order. Dates can be ordered in terms of before or after, or the number of days from the current date. Distances between different locations can be computed using such indicators as the user’s current location and postal code in an address.
Certain kinds of text values can also be ranked. The easiest items to rank are those with where the values fall within a hierarchy or along a scale.
Ordinal values are text values ranked on a scale, for example:
- Excellent is higher than very good
- Beginner is lower than advanced
Taxonomy terms belong to a hierarchy, which provides another means of ranking.
- Items with the same taxonomy term (siblings) are closely related.
- Items that have terms that share the same parent are somewhat related.
- Items that don’t share a common parent aren’t closely related.
The many ways to rank content
Ranking lists provide readers with a familiar way to access detailed information. We have covered common ranking approaches. The possibilities to rank items are extensive. Readers appreciate lists ranking what’s “best”: the best hospitals, the best places to invest, or the best athletes in a given sport. Such lists consider many elements to deliver a ranking. By describing different aspects, a content type’s profile can rank content items in innovative ways so that readers can access the most interesting content.