Structured Product Data

Structured Product Data makes it much easier for customers to both find and understand the products on your website, and can also generate more powerful keywords.

Structuring Product Data at its simplest means supplementing the product name and description with all of its features and benefits specified separately in a standardized and (ideally) normalized way. Features and benefits are expressed as a series of attribute-value pairs; e.g. the product name Degree Women Antiperspirant & Deodorant Invisible Solid Shower Clean 2.6 oz. has the following pairs:

Breaking apart a product name or description into Attribute-Value pairs increases Findability substantially. If you have a Faceted Navigation system (usually employed as a sidebar) customers can “drill down” to the products they want using multiple attributes/facets. All too often, many websites that have invested in Faceted Navigation only provide a few facets. Frequently it’s just a single facet: Brand! Which often leaves customers with an overwhelming number of products and no ability to refine their choices further.

Even if you don’t have Faceted Navigation, you can display Attribute-Value pairs in product details and also use these pairs to generate keyword combinations. These keywords can be further enhanced by giving them a priority weighting.

Attribute-Value Pairs

Attribute-Value pairs improve Understandability as well. Except when customers have an exact specification of what they want (specific products or features), they need help in making a purchase decision. Being able to see the differences between products is an enormous help to Understandability because comparison is fundamental to understanding the trade-offs between features/benefits and price.

Attribute-Value pairs can also improve Understandability in another way. In the deodorant example above, the attributes and attribute values are easy to understand because we’re so familiar with them – or we can easily infer meaning (e.g. “Shower Clean”). But with products that are new to us or we buy infrequently, it’s hard to understand the “so whats” (i.e. benefits) of the features. For example, how many of us who aren’t computer geeks understand the benefits of 2.0 versus 2.6 GHz, quad core versus dual core, 7200 RPM versus 5400 RPM? Websites selling computers are usually in the 5% that have decent Findability because they have Attributes & Values broken out as pairs. But they’re way down the list in terms of Understandability. Unless it’s completely obvious (e.g. Brand, Scent) the “so what” of every Attribute should ideally be explained with the same applying to the Attribute Values.

Standardization, Normalization, and Extension

Beyond the basics of splitting product names into Attribute-Value pairs, your site’s Findability can be supercharged through a combination of Standardization, Normalization, and Extension.

Standardization

Standardization frequently applies to Attribute Values that consist of a number and a unit-of measure. It can also apply to Attributes themselves. For example, here are a few Attribute-Value pairs from camera lenses (note the typo).

Normalization

Normalization is the process of making different but equivalent terms the same. This is critical to Findability and to product comparisons – otherwise real differences can be overwhelmed by apparent differences.

 Examples of different but equivalent terms are:

• Terminology. For Digital Cameras, one manufacturer uses the term “Image Stabilization;” another uses “Vibration Reduction.” For computers, Notebooks are also called Laptops.
• Trademarks. Pull-Ups (Huggies) and Easy-Ups (Pampers) are both trademarks that represent essentially the same feature: training pants.
  

Extension

[IMAGE] Extension means adding more Attribute-Value pairs to the ones derived from the product name. Extended Attribute-Values can boost your Findability substantially, particularly with customers that have specific needs.

Continuing our deodorant example, some people prefer not to use Aluminum based deodorants so extracting the active ingredient (Aluminum Zirconium Tetrachlorohydrex) and Standardizing it as “Aluminum Hydrex” would go a long way to getting high conversion rates from these customers.

Another example is using ingredients to infer which deodorants are aerosols (butane or propane indicates a propellant) so that customers wanting aerosols can filter (or search) by that Attribute Value.