Wouldn’t you like to be able to answer “why” someone liked something, rather than simply “what” they liked? Product attribution personalizes product recommendations versus just product names or SKUs. But product attribution is good for a lot more than just personalization, and today we’ll take a little look at how attributes can be used to create a much better user experience.
The first obvious use for product attributes is for personalization. If you have robust attributes (also referred to as metadata) attached to your products, you can find out why someone likes something. Take, for instance, someone’s smartphone purchases. Product attributes can help answer the question “why did someone buy this phone?” Was it because of the operating system, the manufacturer, the carrier, because it had a physical keyboard (or didn’t), or some other reason? By tracking the attributes of the products the user has either browsed or bought before, you can quickly develop a fingerprint of attributes that define their interest in the product. This can help you recommend other products to them, by matching the highly recurring metadata with new products when they come out. Do this level of personalization, and make sure you have a well-thought-out taxonomy and ontology.
You can easily simplify the browsing process through the use of metadata. This is fairly commonplace now. But when online stores started out, every company’s category structure was needlessly complex. One client we had used the following categories for necklaces: Necklaces > Gold > 24k > 10-12″, and then (because different people browse different ways) also had these categories: Necklaces > 10-12″ > Gold > 24k and Gold > 24k > Necklaces. Obviously, this kind of architecture is a mess, and prone to lots of human error. In fact, there were so many products that were either misplaced or didn’t show up in all the right categories that it made maintaining this site a nightmare. This is an obvious use for metadata. Now this store (and many others) simply has a singular “Necklaces” category and allows the user to sort by product attributes such as “Metal” and “Length.”
Browsing isn’t the only experience that can be improved this way. Searching is arguably more important these days, and using product attributes to aid searches is paramount. In the browsing scenario, the system has the benefit of knowing what level in the product hierarchy to display. Assuming a well-constructed ontology, the system knows which attributes to display and which not to. For example, the “Necklaces” category from the above example used “Length” and “Metal” to assist the user in refining her choice. The system decided not to display “Shoe Size” or other irrelevant attributes because these have nothing to do with necklaces. In the search world, however, the system has to make some educated guesses about what attributes to show or hide. A well-constructed ontology, thesaurus, and user experience goes a long way here as well, and can make or break the search experience.
One of the most powerful uses of metadata is in the world of gift giving. Imagine the clueless husband in a jewelry store (not difficult to imagine). He wants to buy something for his wife because it’s their anniversary, but he’s clueless when it comes to jewelry and what she will like. Knowing this, the salesperson takes a different tact, and asks him what jewelry looks like things she already owns. The guy points out a couple earrings that look familiar, a similar necklace or two, and a bracelet she owns. The salesman then says, “OK, your wife wears white gold jewelry and her favorite stone is amethyst, and she prefers a hoop earring over other styles. Here is a set of hoop earrings that are amethyst set in white gold that she will probably love.”
After reviewing just a few other products that she likes, the salesperson was easily able to identify the product attributes that most of the products had in common. He was then able to search for new products that had similar attributes.
This is a powerful use of product attributes, and one that is not used much today.
Product attributes can also be reverse appended. Doing this lets you ask a very different question. Instead of asking “what types of products does this person like?” we can ask “what types of people buy this product?” If you have a robust system of user attributes, you can attach this data to the products people buy and get a very detailed sense of the types of people interested in each product.
Product attributes can also be used in mass customization as a helpful way to spot trends and guide future product development. Take, for example, a “customize your shoe” feature. Many major shoe brands have such a tool on their site. Imagine if a byproduct of this feature was a report that showed that in the last four months most people have chosen high-top sneakers and pink shoelaces. While it would probably never happen, this could indicate to the brand manager that the ’80s were coming back, and some retro ’80s sneakers might be perfect for next season’s product lineup.
The uses for metadata and product attributes are endless. But, they require a very well-thought-out dictionary, taxonomy, and ontology. Once the hard work in creating these is done, however, you can create rich customer experiences that will put the right products in front of your customers and keep them coming back for more.
Until next time…
Sponsored content in collaboration with Marchex. When it comes to brand keyword bidding, most tests show that it makes sense to bid ... read more
Emily Ma, product director of Tencent’s advertising platform products department, was a keynote speaker at ClickZ Live Shanghai where she discussed the ... read more
The terms that customers type into your site search function can help you to gain an understanding of user behaviour and can be used to optimise ... read more