01

With so much information available on the web, and even on individual web sites, it is important to have access to simple, powerful, and effective search. (Google has made a pretty good business in this domain).

A search engine’s utility can be dramatically improved by using information about the person doing the search, and about the context in which that person is searching. This leaves everyone happier.

As an exercise, let’s consider one day’s worth of search results for one of our customers, an online clothing retailer. Two things immediately become apparent:

One, people are searching very generically. There are searches like “levis skinny jeans men”, but much more common are searches like “shirt”, “blazer”, and “shoes”.

Two, people make a lot of mistakes. Typos abound: “shurt”, “swaroski”, “sleev”, six different misspellings of a particular brand of clothing. Confusion about the clothing retailer’s domain is also

evident: “fishing rod”, “16gb pen drives”, “fort collins”.

This is where context and personalization become very powerful. The more you know about a potential customer, the easier it becomes, when presented with a search like “shirt”, to return a list of shirts that are interesting to that person. If that person is a woman, you’ve just eliminated roughly half your product catalog. If over the course of her visit to your site she’s mostly clicked on clothing in cool colors, you can tilt your results in that direction. If she’s tweeted about how she loves her new sweater from the confusingly-spelled clothier, you might be more likely to include their products in the search results. If in the past she’s bought predominantly wool clothing, this is useful to know – especially if the one time she bought something in cotton she returned it.

Context also makes it much easier to fix errors, or at least to take a good shot at converting them into sales. There are many misspellings of “shirt” for which Google search results would be unlikely to inspire anyone to continue exploring a web site. On a clothing site, however, one might guess that the user intended to search for “shirt”, and respond appropriately. A search for “fishing rod” might return clothing with a fishing-related theme, if there is any. Even if not, the profile of a person interested in a fishing rod might be expected to deviate in subtle but real ways from the profile of a

person interested in a 16gb pen drive. Tell each person that you don’t sell that sort of product, but also offer a selection of clothing that might appeal to each, just in case!

To learn more about our products and services in catalog enrichment, search, and user analytics,
[email protected]

How AI helped in completion of Beethoven’s unfinished 10th symphony

Jeff Jarvis is a US based journalist, professor, and an investor for over five decades. On his blog, Buzzmachine.com, the NYC insider tracks new media developments.
October 19, 2021
know more

Delving Into The Wonder That is 3D Realism

Jeff Jarvis is a US based journalist, professor, and an investor for over five decades. On his blog, Buzzmachine.com, the NYC insider tracks new media developments.
October 17, 2021
know more

The Story of Data with Kuntal Malia

Jeff Jarvis is a US based journalist, professor, and an investor for over five decades. On his blog, Buzzmachine.com, the NYC insider tracks new media developments.
October 9, 2021
know more

AI helps in identifying those at risk of Sleep Apnea

Jeff Jarvis is a US based journalist, professor, and an investor for over five decades. On his blog, Buzzmachine.com, the NYC insider tracks new media developments.
October 7, 2021
know more

Recommended Posts

No comment yet, add your voice below!


Add a Comment

Your email address will not be published. Required fields are marked *