Blog, Semantic Intelligence

One box search? Go semantics!

I have been observing with interest a new trend that is finally emerging among vertical search engines. In order to improve their users’ experience and differentiate themselves from the field crowded with players offering similar services and features, search sites like Mobissimo, LastMinute.com, Dot Homes or service sites like Presdo.com have launched a new feature: a simple one box search to allow consumers to enter in one box all their requirements about a vacation, a home, an auto, etc.

I believe that this is an extremely powerful feature for different reasons.

  • The most obvious is that it can streamline and simplify a search process that unfortunately is still pretty cumbersome. For example, making a reservation for a trip still requires several steps including navigating and entering information in several boxes and sorting and filtering the results in “pre defined” different, but limited and possibly irrelevant for the user, ways to find the best option. Referring to travel websites, when I go to London I generally like to make reservations in hotels in the Earl’s Court area (I like the restaurant selection in the area) with breakfast not included but with free internet access. This, in a traditional site like expedia means 4-5 boxes plus 2 filters and often the information whether internet is free is not available as a filter criteria.
  • In addition, an effective one box search could significantly extend the usability of the site to mobile platforms. Even considering that websites are definitely easier to access in new generation smartphones (like iPhone) the experience is far from being ideal and it is not really addressing the requirements of people that are accessing the site while on the go, like lack of time, not ideal situations to access a site, etc. An effective one box search would allow users to express their needs as easily as sending an SMS or writing a twitter update returning only those entries that really address the user’s needs in order to help him/her jump immediately to the purchase.
  • Lastly, the technology that enables the effectiveness of one box search could also enable the expansion of the search and filtering criteria for the user by mining directly from the free text description of a hotel, the extra elements that could be relevant for the user (see my free internet example above) but that are not part of the standard set of criteria made available on the web site. For example, the extraction of elements, like the availability of a playground or the size or the quality of the illumination of a parking lot, or any other element that could make the difference for many potential customers, could be made available together with the standard search and filtering criteria.

The success of the one box search features is strictly linked to how well the system performs in terms of understanding the request without requiring the user to follow a rigid syntax. From this point of view the sites I have tested are still falling short from what, in my opinion, is the minimum requirement to drive up significantly adoption. For example, Mobissimo seems to have a hard time recognizing New York as a starting location for an air trip (gave me a warning message saying there were too many locations starting with New) but works perfectly if I use LGA or JFK, while LastMinute.com has implemented a mechanism to lead the user in using a syntax making the search more cumbersome than what it should be, and dot homes was not able to disambiguate between Belmont as a road and Belmont as a city.

This is another practical example of where solid semantic technologies can make a difference. Publishers should not ignore that this is not a trivial issue and that cutting corners by implementing internally developed tools or leveraging standard non semantic technologies are workarounds that don’t pay off because they can’t address the complex issues and get the real value that one box search can provide.

 

Author: Luca Scagliarini


Share On

Menu